Bureaucracy as Liability: Administrative Complexity and Polity Collapse across 5,000 Years

Bureaucracy as Liability: Administrative Complexity and Polity Collapse across 5,000 Years | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Bureaucracy as Liability: Administrative Complexity and Polity Collapse across 5,000 Years Ali Raza Jatoi This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9259328/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Does bureaucratic complexity extend or shorten the lifespan of political entities? Classical state formation theory — from Weber to Fukuyama — predicts that administrative depth stabilizes polities by enabling taxation, law enforcement, and territorial control. This paper subjects this claim to its first large-scale empirical test, analyzing 372 polities across 3,400 BCE to 1987 CE using the Seshat Global History Databank and Cliopatria geospatial dataset. Employing Cox proportional hazards models with controls for historical era, world region, and administrative scale, we find that bureaucratic complexity is associated with a 146% increase in collapse hazard (p<0.001) — the opposite of the conventional prediction. Three additional findings sharpen this result. First, the effect decays monotonically across historical eras, remaining strongest in ancient polities and disappearing entirely after 1500 CE. Second, decomposing the bureaucratic index reveals that administrative officials increase collapse risk while professional judicial institutions significantly reduce it. Third, results are robust to restriction of the sample to Eurasia. These findings challenge Weberian assumptions about administrative rationalization as a stabilizing force and identify judicial institutionalization as a neglected mechanism of long-run state durability. state formation bureaucracy polity survival survival analysis cliometrics Seshat Cliopatria Figures Figure 1 Figure 2 Figure 3 I. Introduction In second-century CE Rome, the imperial bureaucracy had never been larger. Salaried officials administered provinces from Britain to Mesopotamia, professional judges adjudicated disputes from codified law, and a courier network connected the capital to its furthest frontiers. By any Weberian measure, the Roman state had achieved administrative maturity. Within three centuries the western empire had ceased to exist. This is not an isolated paradox. The Tang Dynasty reached its bureaucratic apex complete with examination-recruited officials, merit-based promotion, and a sophisticated fiscal apparatus shortly before the An Lushan Rebellion of 755 CE fractured the empire and killed an estimated one-sixth of the world's population. The Abbasid Caliphate developed one of the medieval world's most elaborate administrative systems before fragmenting under the weight of competing factions and fiscal overextension. The Ottoman Empire's famous bureaucratic apparatus, the Sublime Porte, grew most complex precisely during the centuries of territorial contraction that earned it the epithet 'the sick man of Europe.' Across vastly different geographies, cultures, and centuries, a disquieting pattern recurs: states that built the most sophisticated administrative structures did not always survive the longest. Whether this pattern reflects genuine causation, common confounding, or historical coincidence is a question that has animated political theorists for centuries but has never been subjected to systematic quantitative test. Ibn Khaldun observed in 1377 that dynastic bureaucratization accompanied the erosion of social cohesion rather than reinforcing it. Weber argued in 1922 that rational-legal bureaucracy was the foundation of modern state stability. Tainter demonstrated in 1988 that complexity generates diminishing returns until collapse becomes the rational response. Turchin showed in 2009 that elite overproduction within administrative structures drives secular cycles of political crisis. These theorists disagree fundamentally — yet until now no study has tested their competing predictions empirically across the full span of recorded history. This paper does so. Using the Seshat Global History Databank (Turchin et al. 2015) and the Cliopatria geospatial dataset (2025, we apply Cox proportional hazards survival models to 372 polity-phases spanning 3,400 BCE to 1987 CE. We ask a single, precise question: is bureaucratic complexity associated with longer or shorter polity survival, and does this relationship vary across historical eras and types of bureaucratic institution? Three findings emerge. First, higher bureaucratic complexity is associated with significantly greater collapse hazard across the full sample, after controlling for historical era, world region, and polity scale. This result was robust to multiple specifications and confirmed in the Eurasia subsample. Second, this association decays monotonically across historical periods, from its strongest expression in the ancient world to statistical insignificance after 1500 CE. Third, decomposing the bureaucratic index reveals that administrative officials and judicial institutions are associated with collapse hazard in sharply opposing directions: fulltime bureaucrats increase collapse risk while professional judges reduce it. Two important limitations frame everything that follows. The observational design cannot resolve whether bureaucratic complexity causes collapse or results from the political stress that precedes it. Additionally, the Seshat sample is purposive rather than random. With these caveats stated clearly, the patterns that emerge are nonetheless striking in their consistency and theoretical resonance. The paper proceeds as follows. Section II reviews the theoretical literature. Section III describes the data and variables. Section IV presents the empirical strategy. Section V reports results. Section VI discusses theoretical implications and limitations. Section VII concludes. II. Theoretical Background and Literature Review In classical state theory, bureaucratic administration is seen as the bedrock of modern statehood. Weber (1922) argued that rational-legal bureaucracy with salaried, meritocratic officials, a clear hierarchy, and impersonal rules defines the modern state’s legitimacy and efficiency. Such structures are thought to create predictability and solve collective-action problems, theoretically prolonging stability. Fukuyama (2011, 2014) echoes this insight by proposing that political order rests on three pillars (a strong state, rule of law, and accountability) implying that capable bureaucracy is crucial (he notes that in well-functioning democracies “state, law, and accountability all reinforce one another”). Michael Mann (1986) similarly emphasizes centralized bureaucracy as a key source of state power. In this Weberian tradition, greater administrative depth should enhance government reliability and extend polity lifespan (Weber 1922; Fukuyama 2011; Mann 1986). Other traditions turn the Weberian logic on its head. Charles Tilly (1990) famously quipped that “war made the state, and the state made war,” arguing that external threats force rulers to build tax and administrative capacity. In this view, bureaucracies often originate in fiscal-military necessity. However, the flip side is that large bureaucracies may signal a history of conflict and extraction rather than inherent stability. The fiscal-military state literature (Kennedy 1987; Besley & Persson 2011) emphasizes that states expand administration under war- or crisis-driven fiscal pressures, which can become burdensome. Paul Kennedy (1987) highlights imperial overstretch whereby overextended empires collapse under military and fiscal strain. From this perspective, a heavy bureaucratic apparatus can be a symptom of past wars and revenue demands rather than a guarantor of resilience. As a result, Tilly’s framework suggests that extensive administrative capacity can coincide with state vulnerability, challenging the assumption that it alone ensures endurance (Tilly 1990; Kennedy 1987; Besley & Persson 2011). By contrast, theories of complexity and collapse foresee bureaucracy as a potential liability. Joseph Tainter (1988) argues that as polities grow more complex, the marginal benefits of additional complexity, including administrative layers, diminish while maintenance costs rise. Beyond a point the upkeep of a dense bureaucracy yields declining returns, making collapse more likely. Peter Turchin (2009; 2016) extends this logic via secular-cycle models: demographic and fiscal pressures lead to elite overproduction, in which full-time officials swell the ruling class and intensify competition for limited spoils. Turchin documents that periods of heightened elite competition and internal conflict often peak just before historical crises. This resonates with Ibn Khaldun (1377), who attributed dynastic collapse to eroding “asabiyyah” (social cohesion) as ruling dynasties age and bureaucratic palaces grow bloated. Khaldun observed that as states become wealthy and stratified, the solidarity that empowered earlier success wanes, paving the way for decline. The Tainter–Turchin framework predicts that higher bureaucratic complexity can shorten polity lifespan by generating unsustainable maintenance costs and elite factionalism. Economic historians stress that not all institutions are equal: rule-of-law structures can stabilize states even if bureaucracies overreach. North (1990) argues that credible commitment via institutional constraints for example, impartial courts and laws is crucial for development. Acemoglu, Johnson, and Robinson (2001) similarly find that secure property rights and legal constraints on elites drive prosperity. James C. Scott (1998) warns that high-modernist administrative schemes often fail by ignoring local knowledge, implying that bureaucracy is not automatically beneficial. We therefore distinguish administrative capacity from legal institutionalization. Our analysis separates full-time bureaucracy from formal judicial institutions, anticipating (and later confirming) that rule-of-law mechanisms (e.g. professional judges) promote longevity whereas sheer bureaucratic growth can misfire (North 1990; Acemoglu et al. 2001; Scott 1998). We now turn to data and methods. We employ two new global datasets: the Seshat Global History Databank provides coded measures of bureaucratic institutions, and the Cliopatria geospatial dataset maps worldwide polities from 3400 BCE to 2024 CE. We link Seshat’s administrative variables to Cliopatria’s polity durations and fit Cox proportional-hazards models (event-history methods are standard in political demography) to test whether higher bureaucratic complexity predicts longer or shorter polity survival. Author(s) Year Core claim Relevance to this paper Weber 1922 Modern state defined by rational-legal bureaucracy (rule-bound, meritocratic officials) Bureaucracy is presumed to create stability Fukuyama 2011 Strong state + rule of law + accountability → political order Contemporary version of Weber; implies capable bureaucracy is key Mann 1986 Centralized bureaucracy is a key source of state power Supports idea that bureaucracy underpins state strength Tilly 1990 “War made the state, and the state made war” (war drives fiscal/administrative capacity) Bureaucracy emerges from war pressures, may signal stress Kennedy 1987 “Imperial overstretch”: great powers collapse from fiscal/military overreach Connects war-driven bureaucracy to eventual decline Besley & Persson 2011 War and conflict incentivize building state (tax/administration) Modern fiscal-military state models; larger bureaucracies in conflict contexts Tainter 1988 Complex societies collapse as diminishing returns to complexity set in Bureaucratic complexity eventually burdens societies Turchin 2009, 2016 Secular cycles with elite overproduction lead to instability Predicts bureaucratic elites generate factional competition & collapse Ibn Khaldun 1377 Asabiyyah (social cohesion) declines in empires; bureaucracy erodes solidarity Early theory: dynastic decline tied to overgrown bureaucracy North 1990 Institutions (credible commitment, rule of law) drive growth Highlights difference: legal institutions (judiciary) build stability Acemoglu, Johnson & Robinson 2001 Institutions (property rights) determine economic/political outcomes Reinforces that rule-of-law/tradition matters more than mere admin Scott 1998 “High modernism” warns state planning & bureaucracy can fail by ignoring local knowledge Emphasizes dangers of top-down bureaucratic overreach III. Data IIIa. Seshat Global History Databank The primary source for institutional variables is the Seshat Global History Databank (Turchin et al. 2015; Francois et al. 2016), specifically the Equinox-2020 release (seshatdb/Equinox_Data, commit ed8f570, June 2022, distributed via GitHub under CC BY-NC-SA 4.0). Seshat codes institutional, demographic, and military variables for polities across 30 Natural Geographic Areas (NGAs) such as fixed geographic sampling locations selected to represent major world regions and historical trajectories. Each NGA contains a sequence of polity-phases: discrete political formations occupying that location during a defined time interval. The Equinox-2020 release contains 444 polity-phase records across all NGAs. Following standard Seshat practice, 72 records flagged as duplicates, representing polities that span multiple NGAs and are recorded more than once, were excluded, yielding an analytical sample of 372 unique polity-phases. Bureaucratic complexity variables are drawn from the TSDat123 worksheet, which records institutional variables at 100-year time slices within each polity-phase. Six variables were extracted: presence of fulltime bureaucrats (FullTBur), examination-based recruitment (ExamSyst), merit-based promotion (MeritProm), codified legal system (LegCode), professional judges (Judge), and professional lawyers (Lawyer). Each variable is coded on a continuous scale from 0 to 1, where 0 indicates absence, 1 indicates confirmed presence, and intermediate values reflect expert-assessed uncertainty. A composite Bureaucratic Depth Index (BDI) was constructed by averaging the six component variables across all available time slices within each polity-phase, producing a single mean score per polity (BDI_mean). Missing values on individual components were excluded listwise; 39 polity-phases with no bureaucratic data on any component were excluded from Cox model estimation, yielding an effective analytical sample of 333 polities for primary specifications. Polity duration is derived from the Cliopatria geospatial dataset (v0.1.3, January 2025, Zenodo doi: 10.5281/zenodo.14714684 ), which maps the territorial extent and temporal span of worldwide political entities from 3,400 BCE to 2024 CE. Each record contains a polity name, territorial polygon, area in km2, start year (FromYear), end year (ToYear), and a SeshatID field linking to the Seshat Databank where applicable. Polity duration in years was computed as: duration = ToYear minus FromYear. Years BCE are represented as negative integers. Cliopatria and Seshat were merged on PolID, the shared polity identifier. The dependent variable in all survival models is polity duration in years, treated as time-to-event data. All 372 polity-phases represent discrete completed historical episodes; the event indicator (collapsed) equals 1 for all observations. The main independent variable is BDI_mean, ranging from 0 to 1. Controls are: Start year (era), AdmLev (administrative hierarchy levels, serving as a polity scale proxy), and WorldRegion (ten-category fixed effect). Table 1 presents descriptive statistics. Polity duration is strongly right-skewed: the mean of 359 years is approximately double the median of 178 years, indicating most polities were short-lived while a small number of ancient formations persisted for millennia. This distributional property motivates the use of survival analysis rather than ordinary least squares regression Variable N Mean SD Min Median Max Duration (years) 372 359.4 600.1 9 178 4,399 BDI mean 333 0.47 0.36 0.00 0.50 1.00 Admin Levels 319 4.15 1.93 1.0 4.0 9.5 Start Year 372 -444.7 2,251.1 -13,600 222.5 1,896 Note: BDI = Bureaucratic Depth Index (mean of six components, scale 0-1). Start Year in integer form (negative = BCE). Admin Levels = administrative hierarchy levels. N varies due to missing values on individual variables. IV. Methods IVa. Survival Analysis and the Cox Proportional Hazards Model We model polity duration using the Cox proportional hazards framework (Cox 1972), the standard approach for time-to-event data in the social sciences (Box-Steffensmeier and Jones 2004). The Cox model estimates the hazard function, the instantaneous rate of collapse at time t conditional on survival to t, as: h(t) = h0(t) x exp(B1*BDI + B2*Start + B3*AdmLev + B4*WorldRegion) where h0(t) is an unspecified baseline hazard function shared across all observations. This semiparametric formulation makes no distributional assumption about the shape of the baseline hazard. Coefficients are reported as log-hazard ratios (coef) and exponentiated hazard ratios (exp(coef)); a hazard ratio above 1.0 indicates increased collapse risk. All models were estimated in R version 4.4.1 using the survival package (Therneau 2024). IVb. Model Specifications Three nested specifications are estimated. Model 1 (Baseline) includes BDI_mean as the sole predictor, establishing the unconditional association. Model 2 (Era Control — Primary Specification) adds polity start year to control for the strong temporal gradient in both bureaucratic complexity and polity survival. This is the preferred specification. Model 3 (Full Controls) further adds administrative hierarchy levels and world region fixed effects. Model fit is assessed using the concordance statistic, with values above 0.70 indicating satisfactory discrimination. IVc. Proportional Hazards Assumption The Cox model requires that the effect of each covariate on the hazard is constant over time. This was tested using Schoenfeld residuals (Schoenfeld 1982) via the cox.zph function. The global test yielded chi-squared = 4.28 (df = 2, p = 0.12); individual tests for BDI_mean (p = 0.12) and Start year (p = 0.25) both retained the null of proportionality. The assumption is satisfied across all primary specifications. IVd. Additional Analyses Three additional analyses are conducted. First, era subgroup analysis divides the full sample into four historical periods, Ancient (Start 1500), and re-estimates Model 1 within each. Second, component decomposition replaces the composite BDI with its six individual components in a single model alongside the era control. Third, a Eurasia robustness check re-estimates the primary specification on the subsample of polities from Eurasian NGAs (n = 245) to address concerns about differential coverage bias. IVe. Limitations of the Research Design Three limitations warrant explicit statement. First, endogeneity: polities under political stress may develop complex bureaucracies as a crisis response rather than bureaucracy contributing to collapse. The current design supports associational claims only; causal interpretation requires an instrumental variable strategy, identified as a priority for future work. Second, non-random sampling: Seshat's 30 NGA locations were purposively selected, and results describe associations within this sample rather than universal laws of state collapse. Third, index construction: the BDI is an unweighted mean of six components; alternative constructions may yield different results. V. Results Va. Descriptive Patterns The analytical sample comprises 372 unique polity-phases spanning 3,400 BCE to 1987 CE, with bureaucratic complexity data available for 333 polities. Table 1 reveals two features that shape all subsequent analysis. First, polity duration is strongly right-skewed, confirming that most polities were short-lived while a small number of ancient formations persisted for millennia, the longest enduring 4,399 years, the shortest nine. This distributional property is visible in Figure 1 and motivates the use of survival analysis. Second, the Bureaucratic Depth Index exhibits substantial variation, ranging from 0 to 1 with a mean of 0.47, indicating genuinely diverse institutional configurations. Regional composition reflects Seshat's purposive sampling structure. Southwest Asia contributes the largest share (n=93), followed by Europe (n=54), South Asia (n=41), and East Asia (n=39). The implications of uneven coverage for generalizability are addressed in the Eurasia robustness check (Section Vf). Vb. Kaplan-Meier Survival Curves Figure 2 presents Kaplan-Meier survival curves stratified by BDI quartile, with 95% confidence intervals. Polities in the lowest bureaucratic complexity quartile (Q1) exhibit substantially longer survival than higher quartiles. The median survival time for Q1 polities is 289 years, compared with 149, 151, and 129 years for Q2, Q3, and Q4 respectively. Mean survival times show an even sharper contrast: 591 years for Q1 against 185, 170, and 141 years for Q2 through Q4. The log-rank test formally confirms that survival distributions differ significantly across the four quartiles (chi-squared = 94.9, df=3, p<2e-16). Examining observed versus expected collapse counts, Q4 polities collapsed approximately 1.85 times more frequently than expected under the null hypothesis (71 observed versus 38.3 expected), while Q1 polities collapsed considerably less frequently than expected (111 observed versus 187.4 expected). These raw patterns should be interpreted cautiously. Low-complexity polities in Q1 are systematically older, concentrated in the ancient and prehistoric periods when bureaucratic apparatus had not yet developed, and their extended survival may reflect era-specific factors unrelated to bureaucratic complexity. The Cox models address era confounding directly by controlling for polity start year. Vc. Cox Proportional Hazards Models — Primary Results Table 2 reports estimate from three nested Cox models. The proportional hazards assumption was verified prior to interpretation (global Schoenfeld test: chi-squared = 4.28, p=0.12). Across all three specifications, the BDI coefficient is positive and statistically significant, indicating an association between higher bureaucratic complexity and greater collapse hazard. In Model 1 (Baseline), the estimated coefficient is 1.442 (hazard ratio = 4.23, 95% CI: 3.04-5.88, p<0.001), indicating a strong unconditional association. In Model 2 (Era Control — Primary Specification), the BDI coefficient reduces substantially to 0.902 (hazard ratio = 2.46, 95% CI: 1.76-3.44, p<0.001), confirming that era confounding was present and that controlling for it is essential. After accounting for historical era, higher bureaucratic complexity remains associated with approximately 146% greater collapse hazard which is the paper's primary finding. Concordance rises from 0.635 to 0.702, indicating that era controls substantially improve model discrimination. In Model 3 (Full Controls), the BDI coefficient reduces modestly to 0.765 (hazard ratio = 2.15, 95% CI: 1.26-3.66, p=0.005), preserved after controlling for polity scale and regional clustering. Administrative hierarchy levels are not independently associated with collapse hazard (p=0.96), confirming that the BDI finding is not simply a proxy for polity size. The coefficient stability across nested specifications, declining from 1.442 to 0.902 to 0.765 as controls are added but remaining significant throughout, is the primary evidence for robustness. These are associations, not causal estimates; the possibility that bureaucratic complexity responds to instability rather than contributing to it cannot be ruled out with this design. Table 2. Cox Proportional Hazards Models: Bureaucratic Complexity and Polity Collapse Model 1 Baseline Model 2 Era Control Model 3 Full Controls Bureaucratic Depth Index 1.442*** (0.167) 0.902*** (0.170) 0.765** (0.272) Era (Start Year) — 0.0004*** (0.00004) 0.0004*** (0.00005) Admin Levels — — 0.002 (0.047) World Region FE No No Yes Concordance 0.635 0.702 0.701 Observations 333 333 303 LR Test 76.1*** (df=1) 193.8*** (df=2) 171.1*** (df=12) Note: Dependent variable: hazard of polity collapse. Coefficients reported as log-hazard ratios with standard errors in parentheses. Reference category for WorldRegion is Africa. Proportional hazards assumption confirmed (global Schoenfeld test p=0.12). *p<0.05; **p<0.01; ***p<0.001. Vd. Era Heterogeneity Table 3 presents Cox estimates stratified by historical era. A pronounced temporal gradient is evident. Among ancient polities (starting before 500 BCE, n=113), the BDI-collapse association is strongest (coef=2.044, p<0.001). The coefficient remains significant but diminished in the classical period (coef=1.072, p<0.01) and the medieval period (coef=0.984, p<0.01). In the modern period (after 1500 CE, n=46), the coefficient is near zero and statistically indistinguishable from null (coef=0.094, p=0.95). Figure 3 plots these coefficients and 95% confidence intervals, making the monotonic decay pattern visually apparent. Table 3. Cox Proportional Hazards Models Stratified by Historical Era Ancient (1500 CE) BDI coefficient 2.044*** 1.072** 0.984** 0.094 Standard error (0.316) (0.359) (0.311) (0.382) Observations 113 63 111 46 Wald test 41.8*** (df=1) 8.9** (df=1) 10.0** (df=1) 0.06 (df=1) Note: Each column reports a separate Cox proportional hazards model estimated on the indicated era subsample. BDI_mean is the sole predictor in each specification. Era boundaries: Ancient = Start 1500 CE. *p<0.05; **p<0.01; ***p<0.001. Ve. Component Decomposition Table 4 reports Cox estimates replacing the composite BDI with its six individual components alongside the era control (n=138 due to listwise deletion; results are exploratory). Two components achieve statistical significance and point in sharply opposing directions. Fulltime bureaucrats (FullTBur) is positively associated with collapse hazard (coef=1.113, hazard ratio=3.042, 95% CI: 1.171-7.907, p=0.022). Polities with professional administrative officials faced approximately three times the collapse hazard of those without. Professional judges (Judge) is negatively associated with collapse hazard (coef=-1.108, hazard ratio=0.330, 95% CI: 0.164-0.666, p=0.002). Polities with formal judicial institutions faced approximately 67% lower collapse hazard, the only component significantly associated with reduced collapse risk. The remaining four components do not achieve individual significance; multicollinearity among components limits the precision of these estimates and null results should not be interpreted as evidence of no effect. The overall model achieves concordance of 0.738, higher than any composite BDI specification, and is highly significant (Wald test p=2x10-12). The opposing signs on FullTBur and Judge indicate that the aggregate positive BDI-collapse association reflects the dominance of administrative institutions in the composite index, while the stabilizing potential of judicial institutions is masked when all components are averaged together. Table 4. Component Decomposition: Individual Bureaucratic Institutions and Collapse Hazard Component Coef Hazard Ratio p-value Fulltime Bureaucrats (FullTBur) 1.113 3.042 0.022 * Exam-based Recruitment (ExamSyst) 0.224 1.251 0.486 Merit Promotion (MeritProm) -0.188 0.829 0.544 Legal Code (LegCode) 0.420 1.521 0.344 Professional Judges (Judge) -1.108 0.330 0.002 ** Lawyers (Lawyer) 0.156 1.169 0.569 Era (Start Year) 0.0004 — <0.001 *** Observations 138 Concordance 0.738 Note: All six BDI components included simultaneously alongside Start year control. N=138 due to listwise deletion on individual component variables. Results should be treated as exploratory. 95% confidence intervals: FullTBur (1.17-7.91); Judge (0.16-0.67). *p<0.05; **p<0.01; ***p<0.001. Vf. Robustness: Eurasia Subsample Table 5 replicates the three primary Cox specifications on the Eurasia subsample (n=245). In the baseline specification the BDI coefficient is 1.568 (p<0.001). In the era-controlled primary specification the coefficient is 0.647 (p=0.004) — positive, significant at the 1% level, and directionally consistent with the full-sample estimate of 0.902. In the full-controls specification the coefficient is 0.570 but falls short of conventional significance (p=0.055), likely reflecting the further reduction to n=230 when missing AdmLev values are excluded. The directional consistency and significance of the primary specification in the Eurasia subsample provides reassurance that differential data coverage is unlikely to account for the main finding. Table 5. Robustness Check: Eurasia Subsample Model 1 Baseline Model 2 Era Control Model 3 Full Controls Bureaucratic Depth Index 1.568*** (0.205) 0.647** (0.224) 0.570 (0.297) Era (Start Year) — 0.0004*** (0.00005) 0.0004*** (0.0001) World Region FE No No Yes Concordance 0.630 0.688 0.681 Observations 245 245 230 Full sample BDI coef (Model 2) — 0.902*** — Note: Sample restricted to polities from European, Southwest Asian, South Asian, East Asian, Central Eurasian, and Southeast Asian NGAs (n=245). Full-sample primary specification coefficient (Model 2) shown for comparison. Standard errors in parentheses. *p<0.05; **p<0.01; ***p<0.001. VI. Discussion VIa. The Paradox of Bureaucratic Complexity The central finding of this paper is that bureaucratic complexity is positively associated with collapse hazard across 372 polities spanning 5,000 years which is a result robust to controls for historical era, world region, and polity scale, and confirmed in the Eurasia subsample. A one-unit increase in the Bureaucratic Depth Index is associated with approximately 146% higher collapse hazard in the primary specification (p < 0.001). This pattern runs counter to the Weberian prediction and is instead consistent with the Tainter-Turchin framework. Several mechanisms are consistent with this association, though the observational design does not permit adjudication between them. Following Tainter (1988), bureaucratic apparatus may generate diminishing marginal returns with each additional layer of administration requires fiscal resources whose productive yield declines over time. Following Turchin (2009, 2016), the expansion of fulltime administrative positions may accelerate elite overproduction, creating more claimants to state resources than the fiscal system can accommodate, intensifying factional competition and political instability. Alternatively, bureaucratic complexity may be endogenous to instability: polities already under fiscal or military stress may develop complex administrative apparatus as a crisis response. Ibn Khaldun's (1377) observation that dynastic bureaucratization accompanies the erosion of social cohesion rather than causing it directly anticipates this third interpretation. The data cannot distinguish between these mechanisms, and all three warrant serious consideration. VIb. The Temporal Gradient The era subgroup analysis reveals a monotonic decay in the BDI-collapse association: strongest in the ancient world (coef = 2.044, p < 0.001), substantial in the classical and medieval periods (coef approximately 1.0, p < 0.01 in both), and indistinguishable from zero in the modern period (coef = 0.094, p = 0.95). In the ancient world, bureaucratic complexity was genuinely rare and expensive as only a small number of polities had developed administrative apparatus, and those that did may have been overextending their fiscal and organizational capacity relative to available surplus. By the early modern period, bureaucratic organization had become a near-universal feature of surviving states as a form of institutional convergence consistent with the diffusion of the Westphalian state model reducing its value as a predictor of differential survival. This interpretation connects to Fukuyama's (2011) argument that modern states combine bureaucratic capacity with rule of law and accountability in ways that pre-modern states did not. VIc. Administration versus Judiciary: The Core Distinction The component decomposition identifies the paper's sharpest theoretical contribution. Fulltime bureaucrats increase collapse hazard approximately threefold (p = 0.022), consistent with both the fiscal overextension mechanism and Turchin's elite overproduction thesis. Professional judges reduce collapse hazard by approximately 67% (p = 0.002), consistent with North's (1990) credible commitment argument i.e. impartial adjudication reduces transaction costs, protects property rights, and constrains arbitrary executive action. Besley and Persson (2011) argue that legal and fiscal capacity develop jointly as complementary state capacities; the present findings suggest they may have divergent effects on survival when they develop separately or unevenly. These component-level findings are explicitly exploratory given the reduced sample size and observational design; they are best understood as motivating hypotheses for future research. VId. Limitations Four limitations constrain interpretation, the most fundamental of which concerns causal identification. The observational design supports associational claims only. Polities under fiscal or military stress may develop complex bureaucracies as a crisis response, what economists call reverse causation, rather than bureaucracy contributing to collapse. A credible solution requires an instrumental variable: a predictor of bureaucratic complexity that affects collapse hazard only through that channel. Two geographic candidates warrant serious development in future work. The first is river proximity. Rivers historically enabled the infrastructure of early bureaucratic states, grain taxation, record storage, courier networks, and population surveillance along irrigated corridors, making proximity to major river systems a plausible predictor of administrative complexity that is prior to and independent of political outcomes. The hydraulic empire literature, anchored by Wittfogel (1957) and developed empirically by Bentzen et al. (2017) on irrigation and autocracy, provides the theoretical foundation for the relevance condition. Descriptive patterns in the present sample are consistent with this logic: polities located in river-proximate Natural Geographic Areas, e.g. Southern Mesopotamia (Tigris-Euphrates), Upper Egypt (Nile), Middle Yellow River Valley, and Middle Ganga, exhibit a mean Bureaucratic Depth Index of 0.52 compared with 0.47 for non-river NGAs, suggesting the first-stage relationship exists in the data. The harder condition is exclusion: river proximity must affect collapse hazard only through bureaucratic development, not directly. The principal threats are that rivers also predict agricultural productivity, trade wealth, and military vulnerability all of which could independently shape survival. Conditional on the world region fixed effects in Model 3, the trade and productivity channels are at least partially absorbed; military vulnerability is less clearly controlled and represents the strongest remaining concern. The second candidate is terrain ruggedness. Flat terrain reduces the cost of administrative reach — road building, census-taking, tax collection — and should therefore predict higher bureaucratic complexity, while being less obviously related to collapse hazard through non-bureaucratic pathways. Nunn and Puga (2012) provide a well-known application of terrain ruggedness as an instrument in a related historical context, and the Cliopatria dataset's territorial polygons could in principle support construction of NGA-level ruggedness measures from standard digital elevation data — geographic infrastructure that already exists in the present dataset, making this extension genuinely feasible rather than aspirational. A two-instrument design using both river proximity and terrain ruggedness would substantially strengthen identification by enabling overidentification tests, and these strategies are identified as the priority extension of this research programme. A second limitation concerns the non-random structure of the Seshat sample. The 30 Natural Geographic Areas were purposively selected to represent historically significant locations, with substantially denser coverage for Eurasia than for sub-Saharan Africa and the Americas. The findings therefore describe associations within this collection of well-documented polities rather than universal laws of state collapse, and the Eurasia robustness check in Section Vf should be understood as a partial rather than complete response to this concern. Third, the polity-phase unit does not fully capture institutional continuity across phase boundaries; polities that reorganized administratively without territorial dissolution may be coded as separate observations. Fourth, the Bureaucratic Depth Index is an unweighted mean of six components; alternative constructions including factor-analytic or theoretically weighted composites may yield quantitatively different results, though the component decomposition in Section Ve suggests the opposing directional effects of administrative and judicial institutions are robust to the aggregation choice. VII. Conclusion This paper has examined the relationship between bureaucratic complexity and polity survival across 372 polity-phases spanning 3,400 BCE to 1987 CE, using Cox proportional hazards models applied to the Seshat Global History Databank and Cliopatria geospatial dataset. Three principal findings emerge. First, bureaucratic complexity is positively associated with collapse hazard in the full sample after controlling for historical era, world region, and polity scale. A one-unit increase in the Bureaucratic Depth Index is associated with approximately 146% higher collapse hazard in the primary specification robust to alternative model specifications and confirmed in the Eurasia subsample. This association is inconsistent with a simple stabilizing interpretation of pre-modern bureaucracy and is more consistent with the Tainter-Turchin framework of complexity-driven fragility and elite competition. Second, the association between bureaucratic complexity and collapse hazard decays monotonically across historical eras, from its strongest expression in the ancient world to statistical insignificance after 1500 CE. This temporal gradient is consistent with bureaucracy transitioning from an unusual and fiscally demanding institutional form associated with overextension, to a universal baseline feature of surviving modern states among which it no longer differentiates survivors from failures. Third, decomposing the bureaucratic index reveals that administrative officials and judicial institutions are associated with collapse hazard in sharply opposing directions. Fulltime bureaucrats increase collapse risk while professional judges reduce it, a divergence consistent with the rule-of-law literature's distinction between administrative capacity and legal institutionalization as distinct dimensions of state development. Future work should pursue an instrumental variable strategy, river proximity as a predictor of bureaucratic development is one candidate, to move from association to causal identification. Extension to post-1987 polities using V-Dem institutional data would permit examination of whether the null modern-era finding reflects genuine institutional convergence or a data artifact. The central contribution is straightforward: across the longest temporal and geographic scope yet examined quantitatively, bureaucratic complexity is not associated with longer polity survival. The institutions most consistently associated with durability are not administrative but judicial. Whether this reflects a causal mechanism or a correlated signal of deeper institutional quality is a question the present data cannot resolve but one that the findings make considerably more worth asking. Declarations Funding statement: This research received no funding. Author Contribution Ali Raza Jatoi wrote and researched the main manuscript. References Acemoglu, D., Johnson, S., and Robinson, J. A. (2001). The colonial origins of comparative development: An empirical investigation. American Economic Review, 91(5), 1369-1401. Besley, T., and Persson, T. (2011). Pillars of prosperity: The political economics of development clusters. Princeton University Press. Bockstette, V., Chanda, A., and Putterman, L. (2002). States and markets: The advantage of an early start. Journal of Economic Growth, 7(4), 347-369. Borcan, O., Olsson, O., and Putterman, L. (2018). State history and economic development: Evidence from six millennia. Journal of Economic Growth, 23(1), 1-40. Box-Steffensmeier, J. M., and Jones, B. S. (2004). Event history modeling: A guide for social scientists. Cambridge University Press. Cliopatria (2025). Worldwide polity dataset v0.1.3. Zenodo. doi:10.5281/zenodo.14714684 Cox, D. R. (1972). Regression models and life-tables. Journal of the Royal Statistical Society: Series B, 34(2), 187-202. Francois, P., Manning, J., Whitehouse, H., Brennan, R., Currie, T., Feeney, K., and Turchin, P. (2016). A macroscope for global history: Seshat Global History Databank, a methodological overview. Digital Humanities Quarterly, 10(4). Fukuyama, F. (2011). The origins of political order: From prehuman times to the French Revolution. Farrar, Straus and Giroux. Fukuyama, F. (2014). Political order and political decay: From the industrial revolution to the globalization of democracy. Farrar, Straus and Giroux. Ibn Khaldun (1377/1958). The Muqaddimah: An introduction to history (trans. F. Rosenthal). Princeton University Press. Kennedy, P. (1987). The rise and fall of the great powers: Economic change and military conflict from 1500 to 2000. Random House. Mann, M. (1986). The sources of social power, Volume 1: A history of power from the beginning to AD 1760. Cambridge University Press. North, D. C. (1990). Institutions, institutional change and economic performance. Cambridge University Press. Schoenfeld, D. (1982). Partial residuals for the proportional hazards regression model. Biometrika, 69(1), 239-241. Scott, J. C. (1998). Seeing like a state: How certain schemes to improve the human condition have failed. Yale University Press. Tainter, J. A. (1988). The collapse of complex societies. Cambridge University Press. Therneau, T. M. (2024). A package for survival analysis in R. R package version 3.7-0. https://CRAN.R-project.org/package=survival Tilly, C. (1990). Coercion, capital, and European states, AD 990-1990. Blackwell. Turchin, P. (2009). Long-term population dynamics: Models and patterns. University of Connecticut Press. Turchin, P. (2016). Ages of discord: A structural-demographic analysis of American history. Beresta Books. Turchin, P., Brennan, R., Currie, T., Feeney, K., Francois, P., Hoyer, D., Manning, J., Marciniak, A., Mullins, D., Palmisano, A., Peregrine, P., Turner, E., and Whitehouse, H. (2015). Seshat: The Global History Databank. Cliodynamics, 6(1), 77-107. Weber, M. (1922/1978). Economy and society: An outline of interpretive sociology (trans. G. Roth and C. Wittich). University of California Press. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9259328","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":617313692,"identity":"577a9ac9-653f-44c7-841d-bd5b2ee24084","order_by":0,"name":"Ali Raza Jatoi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA9klEQVRIie3QPQrCMBTA8YYHuhS6dhC9Qr2AF3GJFNyyF/wKFOriARREr6CLc0ogXQQP4KK4OujWoYMvxeLU1lEw/yF5lPwIqWWZTL9Yk3C9gV4EDfQIvJpAQRpILsfPlwpSDEjINdJTDXFCiB5p1ms62zAWg82k78yRpMGhlLiShKuF7YOrGlQMDglbSsLJ4nguvwaJZbsAlrI9JIpxJECictJBQjJvBp2crBXb1hEPCdhUgpcTPma7OtLVpCUS6KohFVQJtkcSV72lncxv5J6N/LaU8pmOp2xzkvElDSqe/84v/ka+itrzWO+9T785bDKZTH/WCyREXLKg8H8uAAAAAElFTkSuQmCC","orcid":"","institution":"","correspondingAuthor":true,"prefix":"","firstName":"Ali","middleName":"Raza","lastName":"Jatoi","suffix":""}],"badges":[],"createdAt":"2026-03-29 14:08:10","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-9259328/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9259328/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106476870,"identity":"39a15c35-cbb3-42aa-a997-93c16e073e10","added_by":"auto","created_at":"2026-04-09 03:27:35","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":78314,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of Polity Duration (Years), Analytical Sample. Note: X-axis on log scale. N=372 polity-phases, 3,400 BCE-1987 CE. Source: Seshat Equinox-2020 merged with Cliopatria v0.1.3.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9259328/v1/8f1fddae5fb3d83645c7151e.png"},{"id":106476871,"identity":"260949ab-3304-4bb3-84b4-41e7aaba883b","added_by":"auto","created_at":"2026-04-09 03:27:35","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":156391,"visible":true,"origin":"","legend":"\u003cp\u003eKaplan-Meier Survival Curves by Bureaucratic Complexity Quartile. Note: Shaded bands indicate 95% confidence intervals. Number at risk table shown below curves. Log-rank test: chi-squared = 94.9, df=3, p\u0026lt;2x10-16. N=333 polities with valid BDI d\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-9259328/v1/1a7c4343a00807c37da55dd5.png"},{"id":106476872,"identity":"2f67dce0-ae62-4af3-b470-3ddb412ec5b6","added_by":"auto","created_at":"2026-04-09 03:27:35","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":134392,"visible":true,"origin":"","legend":"\u003cp\u003eBDI Coefficient and 95% Confidence Intervals by Historical Era. Note: Each point represents the BDI_mean coefficient from a separate Cox proportional hazards model estimated on the indicated era subsample.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-9259328/v1/8ca72633a8441321a8beb629.png"},{"id":106724620,"identity":"72e39220-e25f-4b91-8eae-e9f7afd998ea","added_by":"auto","created_at":"2026-04-12 18:28:55","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1268236,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9259328/v1/1107d1da-0d01-4173-98d4-aceeba772256.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Bureaucracy as Liability: Administrative Complexity and Polity Collapse across 5,000 Years","fulltext":[{"header":"I. Introduction","content":"\u003cp\u003eIn second-century CE Rome, the imperial bureaucracy had never been larger. Salaried officials administered provinces from Britain to Mesopotamia, professional judges adjudicated disputes from codified law, and a courier network connected the capital to its furthest frontiers. By any Weberian measure, the Roman state had achieved administrative maturity. Within three centuries the western empire had ceased to exist.\u003c/p\u003e \u003cp\u003eThis is not an isolated paradox. The Tang Dynasty reached its bureaucratic apex complete with examination-recruited officials, merit-based promotion, and a sophisticated fiscal apparatus shortly before the An Lushan Rebellion of 755 CE fractured the empire and killed an estimated one-sixth of the world's population. The Abbasid Caliphate developed one of the medieval world's most elaborate administrative systems before fragmenting under the weight of competing factions and fiscal overextension. The Ottoman Empire's famous bureaucratic apparatus, the Sublime Porte, grew most complex precisely during the centuries of territorial contraction that earned it the epithet 'the sick man of Europe.' Across vastly different geographies, cultures, and centuries, a disquieting pattern recurs: states that built the most sophisticated administrative structures did not always survive the longest.\u003c/p\u003e \u003cp\u003eWhether this pattern reflects genuine causation, common confounding, or historical coincidence is a question that has animated political theorists for centuries but has never been subjected to systematic quantitative test. Ibn Khaldun observed in 1377 that dynastic bureaucratization accompanied the erosion of social cohesion rather than reinforcing it. Weber argued in 1922 that rational-legal bureaucracy was the foundation of modern state stability. Tainter demonstrated in 1988 that complexity generates diminishing returns until collapse becomes the rational response. Turchin showed in 2009 that elite overproduction within administrative structures drives secular cycles of political crisis. These theorists disagree fundamentally \u0026mdash; yet until now no study has tested their competing predictions empirically across the full span of recorded history.\u003c/p\u003e \u003cp\u003eThis paper does so. Using the Seshat Global History Databank (Turchin et al. 2015) and the Cliopatria geospatial dataset (2025, we apply Cox proportional hazards survival models to 372 polity-phases spanning 3,400 BCE to 1987 CE. We ask a single, precise question: is bureaucratic complexity associated with longer or shorter polity survival, and does this relationship vary across historical eras and types of bureaucratic institution?\u003c/p\u003e \u003cp\u003eThree findings emerge. First, higher bureaucratic complexity is associated with significantly greater collapse hazard across the full sample, after controlling for historical era, world region, and polity scale. This result was robust to multiple specifications and confirmed in the Eurasia subsample. Second, this association decays monotonically across historical periods, from its strongest expression in the ancient world to statistical insignificance after 1500 CE. Third, decomposing the bureaucratic index reveals that administrative officials and judicial institutions are associated with collapse hazard in sharply opposing directions: fulltime bureaucrats increase collapse risk while professional judges reduce it.\u003c/p\u003e \u003cp\u003eTwo important limitations frame everything that follows. The observational design cannot resolve whether bureaucratic complexity causes collapse or results from the political stress that precedes it. Additionally, the Seshat sample is purposive rather than random. With these caveats stated clearly, the patterns that emerge are nonetheless striking in their consistency and theoretical resonance.\u003c/p\u003e \u003cp\u003eThe paper proceeds as follows. Section II reviews the theoretical literature. Section III describes the data and variables. Section IV presents the empirical strategy. Section V reports results. Section VI discusses theoretical implications and limitations. Section VII concludes.\u003c/p\u003e"},{"header":"II. Theoretical Background and Literature Review","content":"\u003cp\u003eIn classical state theory, bureaucratic administration is seen as the bedrock of modern statehood. Weber (1922) argued that rational-legal bureaucracy with salaried, meritocratic officials, a clear hierarchy, and impersonal rules defines the modern state\u0026rsquo;s legitimacy and efficiency. Such structures are thought to create predictability and solve collective-action problems, theoretically prolonging stability. Fukuyama (2011, 2014) echoes this insight by proposing that political order rests on three pillars (a strong state, rule of law, and accountability) implying that capable bureaucracy is crucial (he notes that in well-functioning democracies \u0026ldquo;state, law, and accountability all reinforce one another\u0026rdquo;). Michael Mann (1986) similarly emphasizes centralized bureaucracy as a key source of state power. In this Weberian tradition, greater administrative depth should enhance government reliability and extend polity lifespan (Weber 1922; Fukuyama 2011; Mann 1986).\u003c/p\u003e \u003cp\u003eOther traditions turn the Weberian logic on its head. Charles Tilly (1990) famously quipped that \u0026ldquo;war made the state, and the state made war,\u0026rdquo; arguing that external threats force rulers to build tax and administrative capacity. In this view, bureaucracies often originate in fiscal-military necessity. However, the flip side is that large bureaucracies may signal a history of conflict and extraction rather than inherent stability. The fiscal-military state literature (Kennedy 1987; Besley \u0026amp; Persson 2011) emphasizes that states expand administration under war- or crisis-driven fiscal pressures, which can become burdensome. Paul Kennedy (1987) highlights imperial overstretch whereby overextended empires collapse under military and fiscal strain. From this perspective, a heavy bureaucratic apparatus can be a symptom of past wars and revenue demands rather than a guarantor of resilience. As a result, Tilly\u0026rsquo;s framework suggests that extensive administrative capacity can coincide with state vulnerability, challenging the assumption that it alone ensures endurance (Tilly 1990; Kennedy 1987; Besley \u0026amp; Persson 2011).\u003c/p\u003e \u003cp\u003eBy contrast, theories of complexity and collapse foresee bureaucracy as a potential liability. Joseph Tainter (1988) argues that as polities grow more complex, the marginal benefits of additional complexity, including administrative layers, diminish while maintenance costs rise. Beyond a point the upkeep of a dense bureaucracy yields declining returns, making collapse more likely. Peter Turchin (2009; 2016) extends this logic via secular-cycle models: demographic and fiscal pressures lead to elite overproduction, in which full-time officials swell the ruling class and intensify competition for limited spoils. Turchin documents that periods of heightened elite competition and internal conflict often peak just before historical crises. This resonates with Ibn Khaldun (1377), who attributed dynastic collapse to eroding \u0026ldquo;asabiyyah\u0026rdquo; (social cohesion) as ruling dynasties age and bureaucratic palaces grow bloated. Khaldun observed that as states become wealthy and stratified, the solidarity that empowered earlier success wanes, paving the way for decline. The Tainter\u0026ndash;Turchin framework predicts that higher bureaucratic complexity can shorten polity lifespan by generating unsustainable maintenance costs and elite factionalism.\u003c/p\u003e \u003cp\u003eEconomic historians stress that not all institutions are equal: rule-of-law structures can stabilize states even if bureaucracies overreach. North (1990) argues that credible commitment via institutional constraints for example, impartial courts and laws is crucial for development. Acemoglu, Johnson, and Robinson (2001) similarly find that secure property rights and legal constraints on elites drive prosperity. James C. Scott (1998) warns that high-modernist administrative schemes often fail by ignoring local knowledge, implying that bureaucracy is not automatically beneficial. We therefore distinguish administrative capacity from legal institutionalization. Our analysis separates full-time bureaucracy from formal judicial institutions, anticipating (and later confirming) that rule-of-law mechanisms (e.g. professional judges) promote longevity whereas sheer bureaucratic growth can misfire (North 1990; Acemoglu et al. 2001; Scott 1998).\u003c/p\u003e \u003cp\u003eWe now turn to data and methods. We employ two new global datasets: the Seshat Global History Databank provides coded measures of bureaucratic institutions, and the Cliopatria geospatial dataset maps worldwide polities from 3400 BCE to 2024 CE. We link Seshat\u0026rsquo;s administrative variables to Cliopatria\u0026rsquo;s polity durations and fit Cox proportional-hazards models (event-history methods are standard in political demography) to test whether higher bureaucratic complexity predicts longer or shorter polity survival.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Taba\" border=\"1\"\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAuthor(s)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYear\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCore claim\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRelevance to this paper\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWeber\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1922\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eModern state defined by rational-legal bureaucracy (rule-bound, meritocratic officials)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eBureaucracy is presumed to create stability\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFukuyama\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2011\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eStrong state\u0026thinsp;+\u0026thinsp;rule of law\u0026thinsp;+\u0026thinsp;accountability \u0026rarr; political order\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eContemporary version of Weber; implies capable bureaucracy is key\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMann\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1986\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCentralized bureaucracy is a key source of state power\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSupports idea that bureaucracy underpins state strength\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTilly\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1990\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ldquo;War made the state, and the state made war\u0026rdquo; (war drives fiscal/administrative capacity)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eBureaucracy emerges from war pressures, may signal stress\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKennedy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1987\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ldquo;Imperial overstretch\u0026rdquo;: great powers collapse from fiscal/military overreach\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eConnects war-driven bureaucracy to eventual decline\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBesley \u0026amp; Persson\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2011\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWar and conflict incentivize building state (tax/administration)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eModern fiscal-military state models; larger bureaucracies in conflict contexts\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTainter\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1988\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eComplex societies collapse as diminishing returns to complexity set in\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eBureaucratic complexity eventually burdens societies\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTurchin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2009, 2016\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSecular cycles with elite overproduction lead to instability\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePredicts bureaucratic elites generate factional competition \u0026amp; collapse\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIbn Khaldun\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1377\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAsabiyyah (social cohesion) declines in empires; bureaucracy erodes solidarity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eEarly theory: dynastic decline tied to overgrown bureaucracy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNorth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1990\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eInstitutions (credible commitment, rule of law) drive growth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHighlights difference: legal institutions (judiciary) build stability\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAcemoglu, Johnson \u0026amp; Robinson\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eInstitutions (property rights) determine economic/political outcomes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eReinforces that rule-of-law/tradition matters more than mere admin\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eScott\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1998\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ldquo;High modernism\u0026rdquo; warns state planning \u0026amp; bureaucracy can fail by ignoring local knowledge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eEmphasizes dangers of top-down bureaucratic overreach\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e"},{"header":"III. Data","content":"\u003cp\u003e\u003cstrong\u003eIIIa. Seshat Global History Databank\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe primary source for institutional variables is the Seshat Global History Databank (Turchin et al. 2015; Francois et al. 2016), specifically the Equinox-2020 release (seshatdb/Equinox_Data, commit ed8f570, June 2022, distributed via GitHub under CC BY-NC-SA 4.0). Seshat codes institutional, demographic, and military variables for polities across 30 Natural Geographic Areas (NGAs) such as fixed geographic sampling locations selected to represent major world regions and historical trajectories. Each NGA contains a sequence of polity-phases: discrete political formations occupying that location during a defined time interval.\u003c/p\u003e\n\u003cp\u003eThe Equinox-2020 release contains 444 polity-phase records across all NGAs. Following standard Seshat practice, 72 records flagged as duplicates, representing polities that span multiple NGAs and are recorded more than once, were excluded, yielding an analytical sample of 372 unique polity-phases. Bureaucratic complexity variables are drawn from the TSDat123 worksheet, which records institutional variables at 100-year time slices within each polity-phase. Six variables were extracted: presence of fulltime bureaucrats (FullTBur), examination-based recruitment (ExamSyst), merit-based promotion (MeritProm), codified legal system (LegCode), professional judges (Judge), and professional lawyers (Lawyer). Each variable is coded on a continuous scale from 0 to 1, where 0 indicates absence, 1 indicates confirmed presence, and intermediate values reflect expert-assessed uncertainty.\u003c/p\u003e\n\u003cp\u003eA composite Bureaucratic Depth Index (BDI) was constructed by averaging the six component variables across all available time slices within each polity-phase, producing a single mean score per polity (BDI_mean). Missing values on individual components were excluded listwise; 39 polity-phases with no bureaucratic data on any component were excluded from Cox model estimation, yielding an effective analytical sample of 333 polities for primary specifications.\u003c/p\u003e\n\u003cp\u003ePolity duration is derived from the Cliopatria geospatial dataset (v0.1.3, January 2025, Zenodo doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.5281/zenodo.14714684\u003c/span\u003e\u003c/span\u003e), which maps the territorial extent and temporal span of worldwide political entities from 3,400 BCE to 2024 CE. Each record contains a polity name, territorial polygon, area in km2, start year (FromYear), end year (ToYear), and a SeshatID field linking to the Seshat Databank where applicable. Polity duration in years was computed as: duration\u0026thinsp;=\u0026thinsp;ToYear minus FromYear. Years BCE are represented as negative integers. Cliopatria and Seshat were merged on PolID, the shared polity identifier.\u003c/p\u003e\n\u003cp\u003eThe dependent variable in all survival models is polity duration in years, treated as time-to-event data. All 372 polity-phases represent discrete completed historical episodes; the event indicator (collapsed) equals 1 for all observations. The main independent variable is BDI_mean, ranging from 0 to 1. Controls are: Start year (era), AdmLev (administrative hierarchy levels, serving as a polity scale proxy), and WorldRegion (ten-category fixed effect).\u003c/p\u003e\n\u003cp\u003eTable 1 presents descriptive statistics. Polity duration is strongly right-skewed: the mean of 359 years is approximately double the median of 178 years, indicating most polities were short-lived while a small number of ancient formations persisted for millennia. This distributional property motivates the use of survival analysis rather than ordinary least squares regression\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"624\" class=\"fr-table-selection-hover\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eVariable\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eN\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eMean\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eSD\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eMin\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eMedian\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eMax\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDuration (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e372\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e359.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e600.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e178\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e4,399\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBDI mean\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e333\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.00\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAdmin Levels\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e319\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e4.15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.93\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e4.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e9.5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStart Year\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e372\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-444.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2,251.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-13,600\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e222.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1,896\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\u003cstrong\u003e\u003cem\u003eNote:\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cem\u003eBDI = Bureaucratic Depth Index (mean of six components, scale 0-1). Start Year in integer form (negative = BCE). Admin Levels = administrative hierarchy levels. N varies due to missing values on individual variables.\u003c/em\u003e\n \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003cbr\u003e\u003c/div\u003e\n\u003c/div\u003e"},{"header":"IV. Methods","content":"\u003cp\u003e \u003cb\u003eIVa. Survival Analysis and the Cox Proportional Hazards Model\u003c/b\u003e \u003c/p\u003e \u003cp\u003eWe model polity duration using the Cox proportional hazards framework (Cox 1972), the standard approach for time-to-event data in the social sciences (Box-Steffensmeier and Jones 2004). The Cox model estimates the hazard function, the instantaneous rate of collapse at time t conditional on survival to t, as:\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003e \u003cem\u003eh(t) = h0(t) x exp(B1*BDI\u0026thinsp;+\u0026thinsp;B2*Start\u0026thinsp;+\u0026thinsp;B3*AdmLev\u0026thinsp;+\u0026thinsp;B4*WorldRegion)\u003c/em\u003e \u003c/p\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere h0(t) is an unspecified baseline hazard function shared across all observations. This semiparametric formulation makes no distributional assumption about the shape of the baseline hazard. Coefficients are reported as log-hazard ratios (coef) and exponentiated hazard ratios (exp(coef)); a hazard ratio above 1.0 indicates increased collapse risk. All models were estimated in R version 4.4.1 using the survival package (Therneau 2024).\u003c/p\u003e \u003cp\u003e \u003cb\u003eIVb. Model Specifications\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThree nested specifications are estimated. Model 1 (Baseline) includes BDI_mean as the sole predictor, establishing the unconditional association. Model 2 (Era Control \u0026mdash; Primary Specification) adds polity start year to control for the strong temporal gradient in both bureaucratic complexity and polity survival. This is the preferred specification. Model 3 (Full Controls) further adds administrative hierarchy levels and world region fixed effects. Model fit is assessed using the concordance statistic, with values above 0.70 indicating satisfactory discrimination.\u003c/p\u003e \u003cp\u003e \u003cb\u003eIVc. Proportional Hazards Assumption\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe Cox model requires that the effect of each covariate on the hazard is constant over time. This was tested using Schoenfeld residuals (Schoenfeld 1982) via the cox.zph function. The global test yielded chi-squared\u0026thinsp;=\u0026thinsp;4.28 (df\u0026thinsp;=\u0026thinsp;2, p\u0026thinsp;=\u0026thinsp;0.12); individual tests for BDI_mean (p\u0026thinsp;=\u0026thinsp;0.12) and Start year (p\u0026thinsp;=\u0026thinsp;0.25) both retained the null of proportionality. The assumption is satisfied across all primary specifications.\u003c/p\u003e \u003cp\u003e \u003cb\u003eIVd. Additional Analyses\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThree additional analyses are conducted. First, era subgroup analysis divides the full sample into four historical periods, Ancient (Start \u0026lt; -500), Classical (-500 to 500), Medieval (500 to 1500), and Modern (\u0026gt;\u0026thinsp;1500), and re-estimates Model 1 within each. Second, component decomposition replaces the composite BDI with its six individual components in a single model alongside the era control. Third, a Eurasia robustness check re-estimates the primary specification on the subsample of polities from Eurasian NGAs (n\u0026thinsp;=\u0026thinsp;245) to address concerns about differential coverage bias.\u003c/p\u003e \u003cp\u003e \u003cb\u003eIVe. Limitations of the Research Design\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThree limitations warrant explicit statement. First, endogeneity: polities under political stress may develop complex bureaucracies as a crisis response rather than bureaucracy contributing to collapse. The current design supports associational claims only; causal interpretation requires an instrumental variable strategy, identified as a priority for future work. Second, non-random sampling: Seshat's 30 NGA locations were purposively selected, and results describe associations within this sample rather than universal laws of state collapse. Third, index construction: the BDI is an unweighted mean of six components; alternative constructions may yield different results.\u003c/p\u003e"},{"header":"V. Results","content":"\u003ch2\u003eVa. Descriptive Patterns\u003c/h2\u003e\n\u003cp\u003eThe analytical sample comprises 372 unique polity-phases spanning 3,400 BCE to 1987 CE, with bureaucratic complexity data available for 333 polities. Table 1 reveals two features that shape all subsequent analysis. First, polity duration is strongly right-skewed, confirming that most polities were short-lived while a small number of ancient formations persisted for millennia, the longest enduring 4,399 years, the shortest nine. This distributional property is visible in Figure 1 and motivates the use of survival analysis. Second, the Bureaucratic Depth Index exhibits substantial variation, ranging from 0 to 1 with a mean of 0.47, indicating genuinely diverse institutional configurations.\u003c/p\u003e\n\u003cp\u003eRegional composition reflects Seshat\u0026apos;s purposive sampling structure. Southwest Asia contributes the largest share (n=93), followed by Europe (n=54), South Asia (n=41), and East Asia (n=39). The implications of uneven coverage for generalizability are addressed in the Eurasia robustness check (Section Vf).\u003c/p\u003e\n\u003ch2\u003eVb. Kaplan-Meier Survival Curves\u003c/h2\u003e\n\u003cp\u003eFigure 2 presents Kaplan-Meier survival curves stratified by BDI quartile, with 95% confidence intervals. Polities in the lowest bureaucratic complexity quartile (Q1) exhibit substantially longer survival than higher quartiles. The median survival time for Q1 polities is 289 years, compared with 149, 151, and 129 years for Q2, Q3, and Q4 respectively. Mean survival times show an even sharper contrast: 591 years for Q1 against 185, 170, and 141 years for Q2 through Q4.\u003c/p\u003e\n\u003cp\u003eThe log-rank test formally confirms that survival distributions differ significantly across the four quartiles (chi-squared = 94.9, df=3, p\u0026lt;2e-16). Examining observed versus expected collapse counts, Q4 polities collapsed approximately 1.85 times more frequently than expected under the null hypothesis (71 observed versus 38.3 expected), while Q1 polities collapsed considerably less frequently than expected (111 observed versus 187.4 expected).\u003c/p\u003e\n\u003cp\u003eThese raw patterns should be interpreted cautiously. Low-complexity polities in Q1 are systematically older, concentrated in the ancient and prehistoric periods when bureaucratic apparatus had not yet developed, and their extended survival may reflect era-specific factors unrelated to bureaucratic complexity. The Cox models address era confounding directly by controlling for polity start year.\u003c/p\u003e\n\u003ch2\u003eVc. Cox Proportional Hazards Models \u0026mdash; Primary Results\u003c/h2\u003e\n\u003cp\u003eTable 2 reports estimate from three nested Cox models. The proportional hazards assumption was verified prior to interpretation (global Schoenfeld test: chi-squared = 4.28, p=0.12). Across all three specifications, the BDI coefficient is positive and statistically significant, indicating an association between higher bureaucratic complexity and greater collapse hazard.\u003c/p\u003e\n\u003cp\u003eIn Model 1 (Baseline), the estimated coefficient is 1.442 (hazard ratio = 4.23, 95% CI: 3.04-5.88, p\u0026lt;0.001), indicating a strong unconditional association. In Model 2 (Era Control \u0026mdash; Primary Specification), the BDI coefficient reduces substantially to 0.902 (hazard ratio = 2.46, 95% CI: 1.76-3.44, p\u0026lt;0.001), confirming that era confounding was present and that controlling for it is essential. After accounting for historical era, higher bureaucratic complexity remains associated with approximately 146% greater collapse hazard which is the paper\u0026apos;s primary finding. Concordance rises from 0.635 to 0.702, indicating that era controls substantially improve model discrimination. In Model 3 (Full Controls), the BDI coefficient reduces modestly to 0.765 (hazard ratio = 2.15, 95% CI: 1.26-3.66, p=0.005), preserved after controlling for polity scale and regional clustering. Administrative hierarchy levels are not independently associated with collapse hazard (p=0.96), confirming that the BDI finding is not simply a proxy for polity size.\u003c/p\u003e\n\u003cp\u003eThe coefficient stability across nested specifications, declining from 1.442 to 0.902 to 0.765 as controls are added but remaining significant throughout, is the primary evidence for robustness. These are associations, not causal estimates; the possibility that bureaucratic complexity responds to instability rather than contributing to it cannot be ruled out with this design.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2. Cox Proportional Hazards Models: Bureaucratic Complexity and Polity Collapse\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"624\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eModel 1 Baseline\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eModel 2 Era Control\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eModel 3 Full Controls\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBureaucratic Depth Index\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.442*** (0.167)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.902*** (0.170)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.765** (0.272)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEra (Start Year)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.0004*** (0.00004)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.0004*** (0.00005)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAdmin Levels\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.002 (0.047)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWorld Region FE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eConcordance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.635\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.702\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.701\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eObservations\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e333\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e333\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e303\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLR Test\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e76.1*** (df=1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e193.8*** (df=2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e171.1*** (df=12)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eNote:\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cem\u003eDependent variable: hazard of polity collapse. Coefficients reported as log-hazard ratios with standard errors in parentheses. Reference category for WorldRegion is Africa. Proportional hazards assumption confirmed (global Schoenfeld test p=0.12). *p\u0026lt;0.05; **p\u0026lt;0.01; ***p\u0026lt;0.001.\u003c/em\u003e\u003c/p\u003e\n\u003ch2\u003eVd. Era Heterogeneity\u003c/h2\u003e\n\u003cp\u003eTable 3 presents Cox estimates stratified by historical era. A pronounced temporal gradient is evident. Among ancient polities (starting before 500 BCE, n=113), the BDI-collapse association is strongest (coef=2.044, p\u0026lt;0.001). The coefficient remains significant but diminished in the classical period (coef=1.072, p\u0026lt;0.01) and the medieval period (coef=0.984, p\u0026lt;0.01). In the modern period (after 1500 CE, n=46), the coefficient is near zero and statistically indistinguishable from null (coef=0.094, p=0.95). Figure 3 plots these coefficients and 95% confidence intervals, making the monotonic decay pattern visually apparent.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3. Cox Proportional Hazards Models Stratified by Historical Era\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"624\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eAncient (\u0026lt;500 BCE)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eClassical (500 BCE\u0026ndash;500 CE)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eMedieval (500\u0026ndash;1500 CE)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eModern (\u0026gt;1500 CE)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBDI coefficient\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.044***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.072**\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.984**\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.094\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStandard error\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e(0.316)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e(0.359)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e(0.311)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e(0.382)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eObservations\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e113\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e111\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e46\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWald test\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e41.8*** (df=1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e8.9** (df=1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e10.0** (df=1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.06 (df=1)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eNote:\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cem\u003eEach column reports a separate Cox proportional hazards model estimated on the indicated era subsample. BDI_mean is the sole predictor in each specification. Era boundaries: Ancient = Start \u0026lt; -500 BCE; Classical = -500 to 500 CE; Medieval = 500 to 1500 CE; Modern = \u0026gt;1500 CE. *p\u0026lt;0.05; **p\u0026lt;0.01; ***p\u0026lt;0.001.\u003c/em\u003e\u003c/p\u003e\n\u003ch2\u003eVe. Component Decomposition\u003c/h2\u003e\n\u003cp\u003eTable 4 reports Cox estimates replacing the composite BDI with its six individual components alongside the era control (n=138 due to listwise deletion; results are exploratory). Two components achieve statistical significance and point in sharply opposing directions.\u003c/p\u003e\n\u003cp\u003eFulltime bureaucrats (FullTBur) is positively associated with collapse hazard (coef=1.113, hazard ratio=3.042, 95% CI: 1.171-7.907, p=0.022). Polities with professional administrative officials faced approximately three times the collapse hazard of those without. Professional judges (Judge) is negatively associated with collapse hazard (coef=-1.108, hazard ratio=0.330, 95% CI: 0.164-0.666, p=0.002). Polities with formal judicial institutions faced approximately 67% lower collapse hazard, the only component significantly associated with reduced collapse risk. The remaining four components do not achieve individual significance; multicollinearity among components limits the precision of these estimates and null results should not be interpreted as evidence of no effect.\u003c/p\u003e\n\u003cp\u003eThe overall model achieves concordance of 0.738, higher than any composite BDI specification, and is highly significant (Wald test p=2x10-12). The opposing signs on FullTBur and Judge indicate that the aggregate positive BDI-collapse association reflects the dominance of administrative institutions in the composite index, while the stabilizing potential of judicial institutions is masked when all components are averaged together.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 4. Component Decomposition: Individual Bureaucratic Institutions and Collapse Hazard\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"624\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eComponent\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eCoef\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eHazard Ratio\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003ep-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFulltime Bureaucrats (FullTBur)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.113\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3.042\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.022 *\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExam-based Recruitment (ExamSyst)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.224\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.251\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.486\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMerit Promotion (MeritProm)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-0.188\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.829\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.544\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLegal Code (LegCode)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.420\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.521\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.344\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eProfessional Judges (Judge)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-1.108\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.330\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.002 **\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLawyers (Lawyer)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.156\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.169\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.569\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEra (Start Year)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.0004\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026lt;0.001 ***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eObservations\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e138\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eConcordance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.738\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eNote:\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cem\u003eAll six BDI components included simultaneously alongside Start year control. N=138 due to listwise deletion on individual component variables. Results should be treated as exploratory. 95% confidence intervals: FullTBur (1.17-7.91); Judge (0.16-0.67). *p\u0026lt;0.05; **p\u0026lt;0.01; ***p\u0026lt;0.001.\u003c/em\u003e\u003c/p\u003e\n\u003ch2\u003eVf. Robustness: Eurasia Subsample\u003c/h2\u003e\n\u003cp\u003eTable 5 replicates the three primary Cox specifications on the Eurasia subsample (n=245). In the baseline specification the BDI coefficient is 1.568 (p\u0026lt;0.001). In the era-controlled primary specification the coefficient is 0.647 (p=0.004) \u0026mdash; positive, significant at the 1% level, and directionally consistent with the full-sample estimate of 0.902. In the full-controls specification the coefficient is 0.570 but falls short of conventional significance (p=0.055), likely reflecting the further reduction to n=230 when missing AdmLev values are excluded. The directional consistency and significance of the primary specification in the Eurasia subsample provides reassurance that differential data coverage is unlikely to account for the main finding.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 5. Robustness Check: Eurasia Subsample\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"624\" class=\"fr-table-selection-hover\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eModel 1 Baseline\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eModel 2 Era Control\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eModel 3 Full Controls\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBureaucratic Depth Index\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.568*** (0.205)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.647** (0.224)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.570 (0.297)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEra (Start Year)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.0004*** (0.00005)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.0004*** (0.0001)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWorld Region FE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eConcordance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.630\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.688\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.681\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eObservations\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e245\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e245\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e230\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFull sample BDI coef (Model 2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.902***\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eNote:\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cem\u003eSample restricted to polities from European, Southwest Asian, South Asian, East Asian, Central Eurasian, and Southeast Asian NGAs (n=245). Full-sample primary specification coefficient (Model 2) shown for comparison. Standard errors in parentheses. *p\u0026lt;0.05; **p\u0026lt;0.01; ***p\u0026lt;0.001.\u003c/em\u003e\u003c/p\u003e"},{"header":"VI. Discussion","content":"\u003cp\u003e \u003cb\u003eVIa. The Paradox of Bureaucratic Complexity\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe central finding of this paper is that bureaucratic complexity is positively associated with collapse hazard across 372 polities spanning 5,000 years which is a result robust to controls for historical era, world region, and polity scale, and confirmed in the Eurasia subsample. A one-unit increase in the Bureaucratic Depth Index is associated with approximately 146% higher collapse hazard in the primary specification (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). This pattern runs counter to the Weberian prediction and is instead consistent with the Tainter-Turchin framework.\u003c/p\u003e \u003cp\u003eSeveral mechanisms are consistent with this association, though the observational design does not permit adjudication between them. Following Tainter (1988), bureaucratic apparatus may generate diminishing marginal returns with each additional layer of administration requires fiscal resources whose productive yield declines over time. Following Turchin (2009, 2016), the expansion of fulltime administrative positions may accelerate elite overproduction, creating more claimants to state resources than the fiscal system can accommodate, intensifying factional competition and political instability. Alternatively, bureaucratic complexity may be endogenous to instability: polities already under fiscal or military stress may develop complex administrative apparatus as a crisis response. Ibn Khaldun's (1377) observation that dynastic bureaucratization accompanies the erosion of social cohesion rather than causing it directly anticipates this third interpretation. The data cannot distinguish between these mechanisms, and all three warrant serious consideration.\u003c/p\u003e \u003cp\u003e \u003cb\u003eVIb. The Temporal Gradient\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe era subgroup analysis reveals a monotonic decay in the BDI-collapse association: strongest in the ancient world (coef\u0026thinsp;=\u0026thinsp;2.044, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), substantial in the classical and medieval periods (coef approximately 1.0, p\u0026thinsp;\u0026lt;\u0026thinsp;0.01 in both), and indistinguishable from zero in the modern period (coef\u0026thinsp;=\u0026thinsp;0.094, p\u0026thinsp;=\u0026thinsp;0.95). In the ancient world, bureaucratic complexity was genuinely rare and expensive as only a small number of polities had developed administrative apparatus, and those that did may have been overextending their fiscal and organizational capacity relative to available surplus. By the early modern period, bureaucratic organization had become a near-universal feature of surviving states as a form of institutional convergence consistent with the diffusion of the Westphalian state model reducing its value as a predictor of differential survival. This interpretation connects to Fukuyama's (2011) argument that modern states combine bureaucratic capacity with rule of law and accountability in ways that pre-modern states did not.\u003c/p\u003e \u003cp\u003e \u003cb\u003eVIc. Administration versus Judiciary: The Core Distinction\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe component decomposition identifies the paper's sharpest theoretical contribution. Fulltime bureaucrats increase collapse hazard approximately threefold (p\u0026thinsp;=\u0026thinsp;0.022), consistent with both the fiscal overextension mechanism and Turchin's elite overproduction thesis. Professional judges reduce collapse hazard by approximately 67% (p\u0026thinsp;=\u0026thinsp;0.002), consistent with North's (1990) credible commitment argument i.e. impartial adjudication reduces transaction costs, protects property rights, and constrains arbitrary executive action. Besley and Persson (2011) argue that legal and fiscal capacity develop jointly as complementary state capacities; the present findings suggest they may have divergent effects on survival when they develop separately or unevenly. These component-level findings are explicitly exploratory given the reduced sample size and observational design; they are best understood as motivating hypotheses for future research.\u003c/p\u003e \u003cp\u003e \u003cb\u003eVId. Limitations\u003c/b\u003e \u003c/p\u003e \u003cp\u003eFour limitations constrain interpretation, the most fundamental of which concerns causal identification. The observational design supports associational claims only. Polities under fiscal or military stress may develop complex bureaucracies as a crisis response, what economists call reverse causation, rather than bureaucracy contributing to collapse. A credible solution requires an instrumental variable: a predictor of bureaucratic complexity that affects collapse hazard only through that channel.\u003c/p\u003e \u003cp\u003eTwo geographic candidates warrant serious development in future work. The first is river proximity. Rivers historically enabled the infrastructure of early bureaucratic states, grain taxation, record storage, courier networks, and population surveillance along irrigated corridors, making proximity to major river systems a plausible predictor of administrative complexity that is prior to and independent of political outcomes. The hydraulic empire literature, anchored by Wittfogel (1957) and developed empirically by Bentzen et al. (2017) on irrigation and autocracy, provides the theoretical foundation for the relevance condition. Descriptive patterns in the present sample are consistent with this logic: polities located in river-proximate Natural Geographic Areas, e.g. Southern Mesopotamia (Tigris-Euphrates), Upper Egypt (Nile), Middle Yellow River Valley, and Middle Ganga, exhibit a mean Bureaucratic Depth Index of 0.52 compared with 0.47 for non-river NGAs, suggesting the first-stage relationship exists in the data. The harder condition is exclusion: river proximity must affect collapse hazard only through bureaucratic development, not directly. The principal threats are that rivers also predict agricultural productivity, trade wealth, and military vulnerability all of which could independently shape survival. Conditional on the world region fixed effects in Model 3, the trade and productivity channels are at least partially absorbed; military vulnerability is less clearly controlled and represents the strongest remaining concern.\u003c/p\u003e \u003cp\u003eThe second candidate is terrain ruggedness. Flat terrain reduces the cost of administrative reach \u0026mdash; road building, census-taking, tax collection \u0026mdash; and should therefore predict higher bureaucratic complexity, while being less obviously related to collapse hazard through non-bureaucratic pathways. Nunn and Puga (2012) provide a well-known application of terrain ruggedness as an instrument in a related historical context, and the Cliopatria dataset's territorial polygons could in principle support construction of NGA-level ruggedness measures from standard digital elevation data \u0026mdash; geographic infrastructure that already exists in the present dataset, making this extension genuinely feasible rather than aspirational. A two-instrument design using both river proximity and terrain ruggedness would substantially strengthen identification by enabling overidentification tests, and these strategies are identified as the priority extension of this research programme.\u003c/p\u003e \u003cp\u003eA second limitation concerns the non-random structure of the Seshat sample. The 30 Natural Geographic Areas were purposively selected to represent historically significant locations, with substantially denser coverage for Eurasia than for sub-Saharan Africa and the Americas. The findings therefore describe associations within this collection of well-documented polities rather than universal laws of state collapse, and the Eurasia robustness check in Section Vf should be understood as a partial rather than complete response to this concern. Third, the polity-phase unit does not fully capture institutional continuity across phase boundaries; polities that reorganized administratively without territorial dissolution may be coded as separate observations. Fourth, the Bureaucratic Depth Index is an unweighted mean of six components; alternative constructions including factor-analytic or theoretically weighted composites may yield quantitatively different results, though the component decomposition in Section Ve suggests the opposing directional effects of administrative and judicial institutions are robust to the aggregation choice.\u003c/p\u003e"},{"header":"VII. Conclusion","content":"\u003cp\u003eThis paper has examined the relationship between bureaucratic complexity and polity survival across 372 polity-phases spanning 3,400 BCE to 1987 CE, using Cox proportional hazards models applied to the Seshat Global History Databank and Cliopatria geospatial dataset. Three principal findings emerge.\u003c/p\u003e \u003cp\u003eFirst, bureaucratic complexity is positively associated with collapse hazard in the full sample after controlling for historical era, world region, and polity scale. A one-unit increase in the Bureaucratic Depth Index is associated with approximately 146% higher collapse hazard in the primary specification robust to alternative model specifications and confirmed in the Eurasia subsample. This association is inconsistent with a simple stabilizing interpretation of pre-modern bureaucracy and is more consistent with the Tainter-Turchin framework of complexity-driven fragility and elite competition.\u003c/p\u003e \u003cp\u003eSecond, the association between bureaucratic complexity and collapse hazard decays monotonically across historical eras, from its strongest expression in the ancient world to statistical insignificance after 1500 CE. This temporal gradient is consistent with bureaucracy transitioning from an unusual and fiscally demanding institutional form associated with overextension, to a universal baseline feature of surviving modern states among which it no longer differentiates survivors from failures.\u003c/p\u003e \u003cp\u003eThird, decomposing the bureaucratic index reveals that administrative officials and judicial institutions are associated with collapse hazard in sharply opposing directions. Fulltime bureaucrats increase collapse risk while professional judges reduce it, a divergence consistent with the rule-of-law literature's distinction between administrative capacity and legal institutionalization as distinct dimensions of state development.\u003c/p\u003e \u003cp\u003eFuture work should pursue an instrumental variable strategy, river proximity as a predictor of bureaucratic development is one candidate, to move from association to causal identification. Extension to post-1987 polities using V-Dem institutional data would permit examination of whether the null modern-era finding reflects genuine institutional convergence or a data artifact.\u003c/p\u003e \u003cp\u003eThe central contribution is straightforward: across the longest temporal and geographic scope yet examined quantitatively, bureaucratic complexity is not associated with longer polity survival. The institutions most consistently associated with durability are not administrative but judicial. Whether this reflects a causal mechanism or a correlated signal of deeper institutional quality is a question the present data cannot resolve but one that the findings make considerably more worth asking.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eFunding statement:\u003c/h2\u003e \u003cp\u003eThis research received no funding.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eAli Raza Jatoi wrote and researched the main manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAcemoglu, D., Johnson, S., and Robinson, J. A. (2001). The colonial origins of comparative development: An empirical investigation. American Economic Review, 91(5), 1369-1401.\u003c/li\u003e\n \u003cli\u003eBesley, T., and Persson, T. (2011). Pillars of prosperity: The political economics of development clusters. Princeton University Press.\u003c/li\u003e\n \u003cli\u003eBockstette, V., Chanda, A., and Putterman, L. (2002). States and markets: The advantage of an early start. Journal of Economic Growth, 7(4), 347-369.\u003c/li\u003e\n \u003cli\u003eBorcan, O., Olsson, O., and Putterman, L. (2018). State history and economic development: Evidence from six millennia. Journal of Economic Growth, 23(1), 1-40.\u003c/li\u003e\n \u003cli\u003eBox-Steffensmeier, J. M., and Jones, B. S. (2004). Event history modeling: A guide for social scientists. Cambridge University Press.\u003c/li\u003e\n \u003cli\u003eCliopatria (2025). Worldwide polity dataset v0.1.3. Zenodo. doi:10.5281/zenodo.14714684\u003c/li\u003e\n \u003cli\u003eCox, D. R. (1972). Regression models and life-tables. Journal of the Royal Statistical Society: Series B, 34(2), 187-202.\u003c/li\u003e\n \u003cli\u003eFrancois, P., Manning, J., Whitehouse, H., Brennan, R., Currie, T., Feeney, K., and Turchin, P. (2016). A macroscope for global history: Seshat Global History Databank, a methodological overview. Digital Humanities Quarterly, 10(4).\u003c/li\u003e\n \u003cli\u003eFukuyama, F. (2011). The origins of political order: From prehuman times to the French Revolution. Farrar, Straus and Giroux.\u003c/li\u003e\n \u003cli\u003eFukuyama, F. (2014). Political order and political decay: From the industrial revolution to the globalization of democracy. Farrar, Straus and Giroux.\u003c/li\u003e\n \u003cli\u003eIbn Khaldun (1377/1958). The Muqaddimah: An introduction to history (trans. F. Rosenthal). Princeton University Press.\u003c/li\u003e\n \u003cli\u003eKennedy, P. (1987). The rise and fall of the great powers: Economic change and military conflict from 1500 to 2000. Random House.\u003c/li\u003e\n \u003cli\u003eMann, M. (1986). The sources of social power, Volume 1: A history of power from the beginning to AD 1760. Cambridge University Press.\u003c/li\u003e\n \u003cli\u003eNorth, D. C. (1990). Institutions, institutional change and economic performance. Cambridge University Press.\u003c/li\u003e\n \u003cli\u003eSchoenfeld, D. (1982). Partial residuals for the proportional hazards regression model. Biometrika, 69(1), 239-241.\u003c/li\u003e\n \u003cli\u003eScott, J. C. (1998). Seeing like a state: How certain schemes to improve the human condition have failed. Yale University Press.\u003c/li\u003e\n \u003cli\u003eTainter, J. A. (1988). The collapse of complex societies. Cambridge University Press.\u003c/li\u003e\n \u003cli\u003eTherneau, T. M. (2024). A package for survival analysis in R. R package version 3.7-0. https://CRAN.R-project.org/package=survival\u003c/li\u003e\n \u003cli\u003eTilly, C. (1990). Coercion, capital, and European states, AD 990-1990. Blackwell.\u003c/li\u003e\n \u003cli\u003eTurchin, P. (2009). Long-term population dynamics: Models and patterns. University of Connecticut Press.\u003c/li\u003e\n \u003cli\u003eTurchin, P. (2016). Ages of discord: A structural-demographic analysis of American history. Beresta Books.\u003c/li\u003e\n \u003cli\u003eTurchin, P., Brennan, R., Currie, T., Feeney, K., Francois, P., Hoyer, D., Manning, J., Marciniak, A., Mullins, D., Palmisano, A., Peregrine, P., Turner, E., and Whitehouse, H. (2015). Seshat: The Global History Databank. Cliodynamics, 6(1), 77-107.\u003c/li\u003e\n \u003cli\u003eWeber, M. (1922/1978). Economy and society: An outline of interpretive sociology (trans. G. Roth and C. Wittich). University of California Press.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"state formation, bureaucracy, polity survival, survival analysis, cliometrics, Seshat, Cliopatria","lastPublishedDoi":"10.21203/rs.3.rs-9259328/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9259328/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eDoes bureaucratic complexity extend or shorten the lifespan of political entities? Classical state formation theory — from Weber to Fukuyama — predicts that administrative depth stabilizes polities by enabling taxation, law enforcement, and territorial control. This paper subjects this claim to its first large-scale empirical test, analyzing 372 polities across 3,400 BCE to 1987 CE using the Seshat Global History Databank and Cliopatria geospatial dataset. Employing Cox proportional hazards models with controls for historical era, world region, and administrative scale, we find that bureaucratic complexity is associated with a 146% increase in collapse hazard (p\u0026lt;0.001) — the opposite of the conventional prediction. Three additional findings sharpen this result. First, the effect decays monotonically across historical eras, remaining strongest in ancient polities and disappearing entirely after 1500 CE. Second, decomposing the bureaucratic index reveals that administrative officials increase collapse risk while professional judicial institutions significantly reduce it. Third, results are robust to restriction of the sample to Eurasia. These findings challenge Weberian assumptions about administrative rationalization as a stabilizing force and identify judicial institutionalization as a neglected mechanism of long-run state durability.\u003c/p\u003e","manuscriptTitle":"Bureaucracy as Liability: Administrative Complexity and Polity Collapse across 5,000 Years","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-09 03:27:27","doi":"10.21203/rs.3.rs-9259328/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"85761848-e20d-4f1c-ab85-11bd487dd2a7","owner":[],"postedDate":"April 9th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-04-10T14:27:15+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-09 03:27:27","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9259328","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9259328","identity":"rs-9259328","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}