The Marxian Value Tensor: Labor-Hours, Not Dollars

The ValueTensor4x3 is Babylon’s foundational economic data structure. It encodes Marx’s reproduction schema as a computable primitive, measured in labor-hours rather than monetary units. This document explains the theoretical reasoning behind the tensor’s structure, why labor-hours matter, and how the four departments map onto the real economy.

Why a Tensor?

Most economic simulations model GDP, wages, and prices — monetary quantities that obscure the social relations they represent. A wage of $30/hour tells you what the market pays, but not what the worker produces. The difference between what a worker produces and what they are paid is surplus value — the fundamental quantity of capitalist exploitation. Monetary values hide this difference by expressing both wages and output in the same unit (dollars), making exploitation appear as mere “profit margin.”

Marx’s innovation in Capital was to measure everything in labor-hours — the time socially necessary to produce commodities. This makes exploitation directly visible: if a worker works 8 hours but is paid wages equivalent to 4 hours of labor-time, the remaining 4 hours are surplus value appropriated by capital.

The ValueTensor4x3 implements this insight as a computational primitive. Each cell contains labor-hours, not dollars. The conversion from monetary values (QCEW wage data) to labor-hours uses the Socially Necessary Labor Time (SNLT) conversion factor — the monetary expression of labor time (MELT) inverted.

The Three Value Components 

Every commodity contains three components of value, corresponding to three types of labor:

Constant Capital (c)

Dead labor — the labor-time embodied in the means of production consumed during production: raw materials, depreciation of machinery, energy, intermediate inputs. This value is transferred to the product but not created anew; it merely reappears.

In BEA national accounts, constant capital corresponds to intermediate inputs — the value of goods and services consumed as inputs to production. The c/v ratio (organic composition of capital) captures how capital-intensive an industry is: high c/v means more machinery relative to workers, low c/v means more labor-intensive.

Variable Capital (v)

Living labor paid — the labor-time equivalent of wages. This is “variable” because it is the only component that can expand value: a worker paid for 4 hours’ labor-time produces more than 4 hours of value. The difference is surplus value.

In QCEW data, variable capital corresponds to total wages and salaries. The tensor converts these monetary wages to labor-hours using the SNLT factor: v_hours = wages × snlt_factor.

Surplus Value (s)

Living labor unpaid — the labor-time the worker performs but is not compensated for. This is the source of profit, rent, interest, and all forms of capitalist income. The exploitation rate e = s/v measures the ratio of unpaid to paid labor.

BEA national accounts provide the basis for computing s/v ratios: s = gross_output − intermediate_inputs − compensation. This is value added minus labor costs — the surplus appropriated by capital.

The Four Departments 

Marx’s reproduction schema in Capital Volume II divides the economy into departments based on the use-value of output — what the products are for, not what industry produces them.

Department I: Means of Production 

Capital goods that enter the production process as inputs: machinery, raw materials, factory buildings, transportation infrastructure, energy. Output from Department I replaces the constant capital consumed across all departments.

NAICS mapping: Mining (21), Utilities (22), Construction (23), heavy Manufacturing (31-33 partial), Wholesale Trade (42 partial).

Economic significance: Department I determines the rate of accumulation. When Department I grows faster than Department II, capital is investing in its own expansion — the accumulation dynamic that drives the reserve army.

Department IIa: Necessary Consumption 

Wage goods consumed by workers: food, basic housing, clothing, healthcare, transportation to work. Output from Department IIa reproduces the labor force — it is what workers buy with their wages.

NAICS mapping: Food manufacturing and retail (31, 44-45 partial), Healthcare (62), basic Transportation (48-49 partial), basic Accommodation (72 partial).

Economic significance: Department IIa determines the subsistence level. When productivity in IIa rises, the value of labor power falls (workers can be reproduced with less labor-time), which increases relative surplus value even without changing wages.

Department IIb: Luxury Consumption 

Goods consumed by the bourgeoisie and labor aristocracy: luxury housing, fine dining, art, yachts, financial services. Output from Department IIb is purchased with surplus value — it does not reproduce labor power.

NAICS mapping: Finance/Insurance (52), Real Estate (53), luxury Retail (44-45 partial), Professional Services (54), Arts/Entertainment (71).

Economic significance: Department IIb is where surplus value is realized as consumption. The ratio of IIb to IIa output reflects the class distribution of consumption — large IIb relative to IIa indicates high inequality.

The Visibility Scalar: g₃₃ 

Leopoldina Fortunati’s The Arcane of Reproduction (1995) argues that capitalist accumulation depends on the systematic invisibility of reproductive labor. The price system treats care work as a “natural” attribute of women and communities rather than as productive labor that creates value.

The tensor encodes this through the visibility scalar g₃₃:

g₃₃ = 1.0: All care work is monetized. Wages are paid for childcare, cooking, cleaning. The full cost of labor-power reproduction is visible to the price system. No shadow subsidy.
g₃₃ = 0.0: All care work is invisible. Domestic labor is entirely unpaid. Capital appropriates the full value of reproductive labor as a free subsidy. Maximum shadow extraction.
g₃₃ = 0.3 (typical US estimate): About 30% of care work is formally monetized (daycare, restaurants, cleaning services). The remaining 70% is unpaid household labor — a shadow subsidy to capital that does not appear in GDP or wage statistics.

The Fortunati exploitation rate includes this shadow subsidy:

\[e_F = \frac{s + \text{shadow\_subsidy}}{v_{\text{monetized}}}\]

This is always higher than the standard exploitation rate (e = s/v) because it counts the unpaid reproductive labor that makes paid labor possible. The difference between e_F and e is the measure of patriarchal extraction — the surplus appropriated through the gendered division of labor.

SNLT and the Transformation Problem 

The Socially Necessary Labor Time (SNLT) conversion factor transforms monetary values (prices) into labor-hours (values). This is the inverse of the Monetary Expression of Labor Time (MELT):

\[\text{SNLT factor} = \frac{1}{\text{MELT}} = \frac{\text{total labor-hours}}{\text{total value added}}\]

Until SNLT calibration is complete, the tensor uses a default factor of 1.0, meaning values are wage-proportional labor-time proxies. This has an important implication:

Ratios are exact: exploitation rate (s/v), organic composition (c/v), and profit rate (s/(c+v)) are all unit-independent — the SNLT factor cancels in the numerator and denominator. These metrics are fully usable without calibration.
Magnitudes are approximate: the absolute number of labor-hours in a tensor cell is only meaningful after SNLT calibration. Until then, the values represent wage-proportional proxies — they preserve relative ordering and derivatives but not absolute scale.

This separation of concerns means the simulation can operate correctly on ratios and trends (which is what the contradiction field system, Volume I mechanisms, and consciousness dynamics need) while deferring the full transformation problem (prices → values → prices of production) to a future specification.

The NAICS-to-Department Mapping 

The mapping from NAICS industry codes to Marxian departments is a theoretical act, not a technical one. The same industry can produce commodities for multiple departments: a food manufacturer produces both wage goods (IIa — basic groceries) and luxury goods (IIb — artisanal cheese). The department mapping encodes an analysis of the social destination of each industry’s output.

The mapping uses allocation weights that sum to 1.0 per NAICS code. Agriculture (NAICS 11) is allocated 62% to Department I (raw materials as means of production), 33% to Department IIa (food as wage goods), and 5% to Department IIb (luxury agricultural products). These weights are configured in YAML, not hardcoded, enabling recalibration as the analysis deepens.

Excluded sectors: Government (NAICS 92) is excluded from the tensor entirely. Government does not operate within the M-C-M’ circuit — it does not produce commodities for exchange. Government wages appear in the tensor as excluded_wages, tracked but not allocated to any department. This is a theoretical claim: government workers produce use-values (roads, defense, education) but not exchange-values within the capitalist circuit. Whether this exclusion is correct is debatable; the simulation makes it configurable.

The BEA Ratio Pipeline 

The Bureau of Economic Analysis provides national-level input-output data that yields the c/v and s/v ratios needed to decompose wages into the three value components:

\[\frac{s}{v} = \frac{\text{gross output} - \text{intermediate inputs} - \text{compensation}}{\text{compensation}}\]

\[\frac{c}{v} = \frac{\text{intermediate inputs}} {\text{compensation}}\]

These ratios are industry-specific (NAICS-level) and time-varying. The hydration pipeline uses a three-tier lookup hierarchy:

BEA empirical data — most specific, interpolated across years within ±5 years
Sector YAML — 2-digit NAICS defaults from configuration
Department defaults — hardcoded fallback per department (e.g., Department I default c/v = 2.0 reflecting high capital intensity of means-of-production industries)

This hierarchy ensures the tensor always produces values (never fails silently) while preferring empirical data when available.

Detroit Through the Tensor 

The tensor reveals the economic structure of the Detroit metropolitan area in labor-time terms:

Wayne County (26163): High variable capital (large workforce), moderate-to-high exploitation rate (manufacturing wages below value produced), positive imperial rent declining over time (eroding labor aristocracy status). Department I dominant (automotive manufacturing = means of production) with growing Department IIa/III (service economy transition).

Oakland County (26125): High variable capital (professional workforce), lower exploitation rate (higher wages relative to value produced), strong positive imperial rent (suburban labor aristocracy intact). Department IIb dominant (finance, real estate, professional services = luxury consumption).

The exploitation gradient from Wayne to Oakland — visible in the tensor’s per-county exploitation rates — captures the fundamental class geography of the Detroit metro area. The tensor makes this computable rather than merely observable: downstream systems (contradiction fields, reserve army, working day classification) read from the tensor to drive their causal mechanisms.