PRIORITY FILING - INITIAL DRAFT
A fully revised and formatted version will follow.
Gregory Caldwell Beier
Susarb LLC, Cambridge, Massachusetts
March 19, 2026
From the Prolegomena to the Unification of Economics, Computer Science, Mathematics, and Physics by Gregory Caldwell Beier (forthcoming 2026).
© 2026 Gregory Caldwell Beier. Licensed under CC BY-NC-ND 4.0.
Solow (1957) decomposed economic growth into contributions from capital accumulation, labor growth, and a residual - total factor productivity - that accounts for the portion of output growth not explained by measurable inputs. The residual is large, persistent, and poorly understood. This paper proposes that the misallocation component of TFP - the output lost when capital flows to the wrong uses - is a compression artifact: it arises because the price system destroys constraint information that would redirect allocation if it were available. Restuccia and Rogerson (2008) estimated that resource misallocation reduces TFP by 30% to 50% in developing economies. Hsieh and Klenow (2009) sharpened the measurement for China and India, finding that equalizing the distribution of marginal revenue products to US levels would boost manufacturing TFP by 30-50% in China and 40-60% in India (and by 86-115% and 100-128% respectively under full equalization of all distortions). The compression framework developed in the companion papers (Beier, 2026c, 2026b) explains why misallocation occurs: when prices compress away information about how individual transactions interact with shared constraints, capital flows toward uses that appear profitable at the scalar level but destroy value through constraint interactions invisible to the participants. Improving the information architecture - making constraint reasoning visible through structured disclosure with track record incentives - reduces the compression, reduces the misallocation, and produces TFP gains. The scope is deliberately limited to the misallocation component of TFP. It claims that the misallocation component, which the empirical literature has shown to be enormous, is traceable to a specific information-theoretic mechanism, and that repairing that mechanism is a policy variable the growth literature has not identified.
Keywords: total factor productivity, Solow residual, misallocation, information architecture, coordination, price compression, growth, disclosure
JEL Codes: O47, O40, D82, D83, E23
The Solow residual is the largest unexplained number in economics. In growth accounting exercises for advanced economies over the postwar period, total factor productivity typically accounts for 40% to 60% of output growth (Easterly and Levine, 2001; Jones, 2016). The residual measures what the production function leaves out - the portion of output that cannot be attributed to measured increases in capital or labor. Abramovitz (1956) called it "a measure of our ignorance," and Solow (1957) acknowledged that his decomposition identified the residual's magnitude without explaining its source.
Sixty-nine years of subsequent research have produced a catalog of correlates rather than a causal account. TFP growth correlates with technological innovation (Romer, 1990), human capital accumulation (Lucas, 1988), institutional quality (North, 1990; Acemoglu and Robinson, 2012), management practices (Bloom et al., 2013), trade openness, financial development, and rule of law. Syverson's (2011) comprehensive survey documents the range of proposed determinants without converging on a unified explanation. The difficulty is not that the correlates are wrong - they are all plausible channels - but that they describe the residual's neighborhood without identifying its mechanism.
This paper proposes a mechanism for one specific component of TFP: the output lost to misallocation of capital and labor across uses. The Hsieh and Klenow (2009) methodology reveals that misallocation of resources across manufacturing establishments in China and India reduces output by 30% to 60% relative to the US benchmark, and Restuccia and Rogerson (2008) estimate that policy distortions generating misallocation can account for large cross-country TFP differences. The misallocation component is not the whole residual, but it is a large and empirically grounded piece of it.
The mechanism is compression. The price system aggregates dispersed knowledge into scalar signals, but as the companion papers establish (Beier, 2026c; 2026b), the aggregation is lossy - it destroys information about how individual transactions interact with shared constraints. When prices carry less constraint information, capital allocation decisions are made on the basis of incomplete signals, and the resulting allocation diverges from the one that full information would produce. The divergence is misallocation, and the misallocation shows up in the growth accounts as lower TFP.
Hayek (1945) described the price system as a mechanism for coordinating dispersed knowledge. The present paper takes Hayek's insight one step further: if coordination depends on the price system's information content, and if the price system's information content depends on the quality of the information architecture feeding it, then the coordination quality of an economy is a measurable property of its information architecture. TFP, to the extent it captures coordination quality, is the output of that architecture. Improve the architecture and you improve the residual.
The paper proceeds as follows. Section 2 argues that TFP is better understood as a measure of coordination quality than as a measure of technology. Section 3 develops the mechanism connecting price compression to misallocation. Section 4 examines historical and cross-country evidence. Section 5 connects the Disclosure Incentive to TFP growth. Section 6 presents testable predictions. Section 7 states limitations. Section 8 concludes.
The standard growth accounting framework decomposes output Y into a production function of capital K and labor L, scaled by total factor productivity A:
Y = A * F(K, L)
Under constant returns to scale and competitive factor markets, A captures everything that affects output beyond the measured quantities of K and L. The conventional interpretation is that A represents "technology" - the stock of knowledge about how to transform inputs into outputs.
The difficulty with this interpretation is that technology is an input. A new production technique, a better machine design, a more efficient chemical process - each of these is a form of knowledge that enters the production function by changing the relationship between inputs and outputs. If technology were fully measured, it would appear inside F rather than in A. The residual captures technology only to the extent that technology is unmeasured, which raises the question of what else A might contain.
An alternative interpretation has been present in the literature since Hayek but never fully formalized: A measures the quality of coordination. Given a fixed stock of capital and labor, output depends on how effectively those inputs are matched to their highest-value uses. A skilled worker misallocated to a low-productivity firm produces less than the same worker at a high-productivity firm. Capital invested in a venture that looks profitable at the price level but interacts destructively with environmental constraints it cannot observe produces less social value than the same capital directed by full constraint information. The gap between actual output and the output that would obtain under perfect allocation is the coordination loss, and it lives in A.
Hall and Jones (1999) approached this from the "social infrastructure" direction, arguing that differences in output per worker across countries are primarily attributable to differences in "social infrastructure" - the institutions and policies that shape economic incentives. Their social infrastructure concept includes property rights enforcement, openness to trade, and government quality. The information architecture this paper identifies is a specific component of social infrastructure: the system through which dispersed knowledge about constraints, opportunities, and risks enters the price system and informs allocation decisions.
North's (1990) institutional economics provides the broader frame. Institutions reduce transaction costs and thereby improve coordination. The information architecture is the subset of institutional infrastructure that determines how much knowledge the price system carries and how accurately it carries it. Acemoglu and Robinson's (2012) distinction between extractive and inclusive institutions maps onto the information architecture framework: extractive institutions suppress information to protect rents (Olson, 1982), while inclusive institutions produce richer information environments that improve the coordination mechanism's reach. The TFP gap between economies with extractive and inclusive institutions is, on this interpretation, substantially an information architecture gap.
None of this is to say that technology, human capital, and institutions as conventionally understood are unimportant. They are. But the channel through which they affect TFP runs at least partly through the information architecture. Romer's (1990) endogenous growth model treats ideas as non-rival inputs that drive growth by expanding the production possibility set. The present paper addresses a different margin: not the production of new ideas but the transmission of existing knowledge through the coordination mechanism. A brilliant innovation that the price system cannot transmit into allocation decisions contributes less to TFP than a modest improvement that the information architecture carries efficiently to every agent who needs it. Lucas's (1988) human capital externality - educated workers improving the productivity of those around them - is itself an information architecture effect: knowledge spillovers are a form of uncompressed information transmission whose value depends on the channels through which the spillover travels. Cross-country evidence from Comin and Hobijn (2010) confirms that technology adoption speed varies enormously across countries; the compression framework suggests that adoption speed is itself partly determined by information architecture quality, since adopting a new technology requires learning about its properties, costs, and constraint interactions - information that must travel through whatever channels the economy provides.
The mechanism connecting information architecture to TFP operates through three links, each grounded in established results.
The first link is that prices compress. Every market price maps the high-dimensional judgments of participating agents into a scalar. Shannon's (1948) channel capacity theorem requires that compressing a signal below its entropy rate destroys information. The companion paper (Beier, 2026c) applies this framework to the price system specifically, establishing that the information destroyed includes judgment about constraint environments - the interactions between individual transactions and shared resources (atmospheric carbon budgets, aquifer depletion, infrastructure capacity, ecosystem resilience) that the scalar price cannot carry.
The second link is that compression causes misallocation. When the price signal is missing constraint information, agents allocating capital on the basis of that signal direct resources toward uses whose profitability is overstated by the compressed price. A factory that appears profitable at current input prices but draws unsustainably on an aquifer it shares with five neighboring operations is an example: the price of water does not reflect the interaction effects, so each factory's individual allocation decision looks rational while the aggregate outcome is destructive. Hsieh and Klenow (2009) did not frame their findings in information-theoretic terms, but their methodology is consistent with the compression framework. They compared the distribution of marginal products of capital and labor across manufacturing plants in China and India to a benchmark (the United States) where prices are presumably less distorted, and found that equalizing marginal products to US-benchmark levels would increase manufacturing TFP by 30-50% in China and 40-60% in India, with gains of 86-115% and 100-128% respectively under the more aggressive full-equalization scenario. The distortions they measured - heterogeneous markups, differential access to credit, regulatory barriers that vary across firms - are, in the compression framework, instances of price signals that fail to carry the information that would direct resources to their highest-value use.
Restuccia and Rogerson (2008) provided the theoretical complement, showing that policy distortions equivalent to heterogeneous taxes and subsidies across establishments generate TFP losses of 30-50% in calibrated models. Their distortions are not explicitly informational, but the mechanism is the same: when the signals guiding allocation diverge from the signals that full information would produce, resources end up in the wrong places and measured TFP falls.
The third link is that repairing the compression reduces misallocation and raises TFP. The Disclosure Incentive (Beier, 2026b) improves the information architecture by introducing a cooperative channel through which structured constraint reasoning enters the price system. When capital allocators can observe disclosed assessments of how firms interact with constraint environments - scored against outcomes and attributed to identifiable analysts - allocation decisions incorporate information that the compressed price alone does not carry. The misallocation attributable to missing constraint information shrinks, and the TFP component that misallocation was dragging down improves.
The claim is deliberately circumscribed. The compression mechanism does not explain the portion of TFP attributable to genuine technological innovation, to learning-by-doing, or to organizational improvements within firms. It explains the portion attributable to misallocation of resources across uses - the portion that arises because the price system's information content is insufficient to guide allocation to its most productive destinations.
The historical record is consistent with the compression framework, though the evidence is suggestive rather than causal.
Major TFP accelerations in the historical record coincide with improvements in the information architecture of economic coordination. The telegraph (1840s-1860s) collapsed the time required for price information to travel across distances from days to seconds, producing what economic historians have documented as a dramatic reduction in price dispersion across geographically separated markets. The introduction of standardized financial disclosure mandates following the Securities Acts of 1933 and 1934 created a common information environment for capital allocation in the United States, and the postwar TFP acceleration in the US coincided with the maturation of this disclosure infrastructure. The Bloomberg terminal (1980s) and the internet (1990s) each expanded access to financial and economic information by orders of magnitude, and each period saw TFP acceleration in the economies most exposed to the new infrastructure.
Conversely, TFP decelerations and collapses coincide with degradation of information architecture. The Soviet economy's persistent TFP underperformance relative to market economies, despite massive capital investment, is consistent with the compression framework: centralized planning suppressed the price system's information content by replacing market prices with administratively determined ones, and the resulting misallocation was enormous. Parente and Prescott (1994) modeled barriers to technology adoption as a source of cross-country TFP differences; in the compression framework, barriers to adoption are partly barriers to information transmission - the knowledge about new technologies, their costs, and their constraint interactions cannot travel through channels that the institutional environment has suppressed.
Financial crises produce TFP declines that the compression framework helps explain. When trust in price signals collapses - as it did in 2008 when the prices of mortgage-backed securities turned out to carry almost no information about the underlying credit quality - the information architecture degrades sharply. Capital allocation becomes defensive, uncertainty premia rise, and resources flow to safe rather than productive uses. The measured TFP decline during and after the 2008 crisis is consistent with a temporary collapse in the information content of the price system.
Cross-country evidence points in the same direction. Economies with stronger press freedom, more developed financial disclosure regimes, greater market transparency, and better-functioning judicial systems - all proxies for information architecture quality - tend to have higher TFP, controlling for capital, labor, human capital, and standard institutional variables. The correlation is well-documented in the institutions-and-growth literature (Hall and Jones, 1999; Acemoglu and Robinson, 2012), though it has not been framed explicitly as an information architecture effect. The contribution of the present paper is to provide the mechanism that connects these institutional proxies to TFP: they all affect the quality, richness, and fidelity of the information environment through which economic agents coordinate.
The Disclosure Incentive (Beier, 2026b) is a specific, implementable mechanism for improving information architecture quality. Its connection to TFP operates through the misallocation channel developed in Section 3.
Under the current information architecture, capital allocators make decisions on the basis of prices, financial statements, ESG ratings, analyst reports, and internal models. Each of these information sources compresses judgment to some degree: prices are scalars, financial statements standardize into templates, ESG ratings diverge across providers because they compress judgment into scores (Berg, Koelbel, and Rigobon, 2022), and analyst reports are filtered through institutional incentives. The compression means that constraint information - how a firm's operations interact with aquifer depletion, atmospheric loading, grid capacity, ecosystem health - is either absent from the allocator's information set or present only in degraded form.
The Disclosure Incentive enriches the information set. When firms' analysts disclose structured constraint reasoning - scored against outcomes and accumulated into track records - the allocator gains access to judgment that was previously compressed away. Capital allocation incorporates the constraint information, and the portion of misallocation attributable to missing constraint information shrinks.
The TFP implication follows. If misallocation is reduced by x% through improved information architecture, and if misallocation accounts for y% of the TFP gap relative to the efficient benchmark, then TFP improves by a fraction of x * y. The Hsieh-Klenow estimates suggest y is large (30-60% of manufacturing output in developing economies, smaller but still substantial in advanced economies). Even modest values of x - even a 10% reduction in the misallocation attributable to missing constraint information - would produce measurable TFP gains at the sector level.
The gains should concentrate in sectors where constraint information is most compressed and most consequential: energy (carbon constraints), agriculture (water and soil constraints), infrastructure (capacity and resilience constraints), and natural resource extraction (depletion constraints). In sectors where the relevant constraints are well-measured and well-priced - commodity markets with transparent spot and futures prices, for instance - the marginal value of additional disclosure is lower, and the TFP gain from improved information architecture is correspondingly smaller.
Prediction 1 (Cross-country TFP and information architecture). An index of information architecture quality - constructed from press freedom scores, financial disclosure standards, market transparency indices, and judicial system quality - should explain cross-country TFP variation with explanatory power equal to or exceeding standard institutional proxies (rule of law, property rights, government effectiveness) after controlling for capital, labor, and human capital.
Identification strategy: Cross-country panel regression with the information architecture index as the key independent variable, controlling for factor inputs and standard institutional measures. The prediction is that the information architecture index retains explanatory power after the standard variables are included.
Prediction 2 (Sector-level TFP after disclosure adoption). Sectors that adopt structured constraint disclosure frameworks should show measurable TFP gains within five to ten years, relative to comparable sectors without adoption. The gains should be largest in constraint-intensive sectors (energy, agriculture, water-dependent manufacturing).
A naive difference-in-differences comparison between adopting and non-adopting sectors is vulnerable to selection bias: sectors that voluntarily adopt structured disclosure likely differ systematically in management quality, institutional governance, and technological sophistication. Clean identification requires either an exogenous policy shock (a regulatory mandate imposing disclosure requirements on specific sub-sectors but not others) or an instrumental variable strategy that isolates the causal effect of information architecture from endogenous sector characteristics. The prediction is testable in principle but demands careful econometric design.
Identification strategy: Difference-in-differences at the sector level, comparing TFP growth in treatment sectors (with disclosure adoption) to control sectors (without), within the same economy and time period.
Prediction 3 (Misallocation reduction). The dispersion of marginal revenue products of capital and labor across establishments within a sector - Hsieh and Klenow's measure of misallocation - should decrease in sectors where structured disclosure improves the information available to capital allocators. The reduction should be measurable as a narrowing of the interquartile range of establishment-level productivity within the sector.
Identification strategy: Compare the distribution of establishment-level productivity measures before and after disclosure adoption, using non-adopting sectors as controls.
Prediction 4 (Crisis resilience). Economies with richer information architectures should experience smaller TFP declines during financial crises, because the information content of their price systems is more resilient to confidence shocks. The prediction is that TFP declines during the next major financial disruption will be negatively correlated with pre-crisis information architecture quality, controlling for the severity of the financial shock.
Identification strategy: Cross-country comparison of TFP trajectories during and after a common financial shock, with pre-crisis information architecture quality as the key explanatory variable.
The paper proposes a framework and a mechanism, not a formal proof. A general equilibrium growth model that derives TFP as a function of information architecture quality - with Shannon compression entering the production function through its effect on allocation efficiency - would be the definitive formalization. That model is beyond the scope of this paper and is identified as a direction for future work.
TFP measurement is notoriously difficult. Growth accounting requires estimates of capital stocks, labor inputs, and factor shares that are themselves subject to measurement error, and the residual inherits all of the errors in the measured inputs. Syverson (2011) surveys the measurement challenges. The predictions in Section 6 are framed in terms of relative changes (treatment vs. control) rather than absolute levels to mitigate this concern, but measurement noise will reduce statistical power.
The causal direction between information architecture and TFP is not established by the historical correlations in Section 4. Richer economies can afford better information infrastructure, creating a reverse causality channel. The instrumental variable strategy for addressing this - using exogenous variation in information architecture quality, such as the introduction of disclosure mandates or the arrival of new communication technologies - is standard in the growth literature but has not been implemented for information architecture specifically.
The paper claims to explain the misallocation component of TFP, not the innovation component. Genuine technological progress - the production of new knowledge about how to transform inputs into outputs - operates through channels that the information architecture framework does not address directly. Romer's (1990) endogenous growth theory and the innovation economics literature remain the relevant frameworks for that portion of the residual. The present paper's contribution is to identify a specific, large, and empirically grounded component of TFP (misallocation) as a compression artifact, and to propose a specific mechanism (the Disclosure Incentive) for reducing it.
The information architecture index proposed in Prediction 1 does not yet exist as a standardized measure. Constructing it from existing proxies (press freedom, disclosure standards, transparency indices) introduces measurement error and conceptual imprecision. A purpose-built measure of information architecture quality - one that directly assesses how much constraint information the price system carries - would strengthen the empirical program but requires data infrastructure that does not yet exist.
Solow's residual measures the gap between what an economy's inputs could produce under perfect coordination and what they actually produce. The gap persists because the price system - the mechanism through which coordination occurs - compresses away constraint information that would improve allocation if it were available. The compressed information is not lost to the economy in the sense that no agent ever possessed it; in many cases the relevant knowledge exists among specialists, engineers, scientists, and analysts who have no channel through which to make it available to the agents whose allocation decisions it would inform.
The Disclosure Incentive provides that channel. Structured, scored, attributed disclosure of constraint reasoning feeds information into the price system that the price system cannot generate on its own. The coordination mechanism improves. Allocation shifts toward uses that account for constraint interactions rather than ignoring them. The misallocation component of TFP shrinks.
The implication for growth policy is direct. The standard growth policy toolkit - invest in physical capital, invest in human capital, strengthen institutions, promote innovation - addresses the inputs to growth and the institutional environment in which growth occurs. The information architecture is a distinct policy variable that operates through a distinct channel: not what the economy knows how to produce, but how effectively the economy coordinates what it already knows. Improving the information architecture does not require new technology or new institutions in the conventional sense. It requires making visible the constraint reasoning that economic agents already possess but currently have no incentive to disclose, no format in which to structure it, and no mechanism through which to have it verified. The Disclosure Incentive provides the incentive, the format, and the mechanism. The TFP gain is the coordination dividend.
This paper presents results from the Prolegomena to the Unification of Economics, Computer Science, Mathematics, and Physics (Beier, forthcoming 2026). The connection between information architecture and total factor productivity builds directly on the compression framework developed in the companion papers on the Grossman-Stiglitz paradox (Beier, 2026d) and the information-theoretic analysis of the price system (Beier, 2026c). The author invites engagement from the growth, macroeconomics, and development economics communities.