A central idea of the Accounting Bias framework is deceptively simple:
\[ AB=A-E, \]
where \(A\) denotes the accounting representation of an economic phenomenon and \(E\) denotes its underlying economic value.
The equation immediately raises a deeper problem.
What exactly is \(E\)?
It is tempting to imagine economic value as an objective quantity existing independently of accounting: somewhere behind the balance sheet lies the "true" value of the firm's assets, liabilities, equity and income, and accounting merely succeeds or fails in discovering it.
But the history of economic thought gives little support to such a simple interpretation.
From Adam Smith and David Ricardo to Karl Marx, the marginalist revolution and modern economic theory, economists have disagreed not merely about how to measure value, but about what value is and what determines it.
Maurice Dobb's Theories of Value and Distribution since Adam Smith: Ideology and Economic Theory is particularly useful in this respect. Dobb presents the development of value theory as a history of competing analytical systems, from Smith and Ricardo through Marx and the marginalist revolution to later debates over capital and distribution.
This history has an important implication for accounting.
If economics itself does not provide a unique, independently observable measure of economic value, accounting cannot be understood simply as an imperfect attempt to record a quantity that economists already know how to measure.
The relationship is more subtle.
Accounting is one measurement system imposed upon an economic reality whose value is itself theory-dependent, uncertain and often fundamentally unobservable.
Accounting measurement begins with quantities that appear concrete:
\[ \text{Assets}, \qquad \text{Liabilities}, \qquad \text{Income}, \qquad \text{Equity}. \]
But these categories quickly lead to questions of economic value.
Suppose a machine appears on a balance sheet at €1 million.
Its accounting carrying amount is observable:
\[ A=1{,}000{,}000. \]
But what is its economic value?
One possibility is historical acquisition cost. Another is replacement cost. Another is the present value of future cash flows generated by the machine. Another is its market selling price. Another is the contribution the machine makes to the value of the productive system in which it operates.
These quantities need not coincide.
The difficulty is therefore not simply that accountants have chosen the wrong number.
Before criticizing accounting measurement, one must specify which economic conception of value should constitute the benchmark.
That question belongs to economic theory.
And economics has never supplied a unique answer.
Classical political economy placed value near the center of economic analysis.
For Adam Smith, the problem already contained distinctions that would remain difficult for subsequent economics.
A commodity could be useful without commanding a high exchange value, while another commodity could possess substantial exchange value despite having a very different relation to human usefulness.
Economic value therefore could not simply be identified with physical usefulness.
The classical problem became one of explaining the systematic relationships according to which commodities exchange.
Production, labour, costs and distribution consequently became central to the theory of value.
The important point for accounting is not whether Smith's particular theory should be adopted today.
It is that value required a theory.
It was not an observable physical property of the commodity.
There is no instrument analogous to a scale or thermometer that reveals the economic value of a machine.
Economic theory must first determine what kind of magnitude is being sought. Only afterward can the question of measurement arise.
Ricardo pushed the classical analysis toward a more systematic theory connecting production, value and distribution.
Within this tradition, value was closely connected to the conditions under which commodities were produced.
The attraction of such an approach is clear.
If value can ultimately be related to objective conditions of production, then it appears possible to ground economic value in something deeper than temporary market prices.
The market price observed today need not equal the underlying value.
One can therefore distinguish between
\[ P_t \]
and some more fundamental magnitude
\[ V. \]
This distinction is extremely important.
It means that observable price and economic value are not necessarily identical.
The problem encountered in accounting therefore already appears inside economic theory itself.
Accounting may provide an observable number. Markets may provide another observable number. Neither automatically establishes the underlying economic benchmark.
Marx developed the classical approach in another direction, maintaining a distinction between observable exchange relations and the social relations underlying them.
Again, whether one accepts Marx's theory is secondary to the methodological point relevant here.
Economic value was not treated simply as whatever number happened to appear in a market transaction.
Price and value could diverge.
Consequently,
\[ P \neq V \]
was not necessarily an anomaly.
The observable market price was itself a representation of deeper economic relations.
This creates an interesting parallel with Accounting Bias.
The accounting framework distinguishes
\[ A \]
from
\[ E. \]
Classical value theory distinguishes, in different ways,
\[ P \]
from the determinants underlying \(P\).
Both reject the proposition that an observable monetary number automatically exhausts the economic phenomenon being represented.
The marginalist revolution transformed the problem.
Value increasingly became connected to scarcity, preferences, utility and marginal choice.
The center of analysis moved away from the classical emphasis on production and distribution toward the relationship between scarce goods and individual economic agents.
This solved some problems.
But from the standpoint of measurement it created another.
Utility is not directly observable.
If an individual's preferences are represented by
\[ U(x_1,\ldots,x_n), \]
the function is an analytical representation of preferences, not a physical quantity that an accountant can observe.
Moreover, in standard consumer theory utility is generally ordinal.
If
\[ U(A)>U(B), \]
the statement represents preference ordering.
It does not imply that the numerical difference
\[ U(A)-U(B) \]
possesses an independently measurable economic meaning.
The marginalist revolution therefore did not transform economic value into an observable physical magnitude.
It transformed the theoretical framework through which value was explained.
This is where Dobb becomes particularly important for the Accounting Bias framework.
His history suggests that theories of value cannot be regarded simply as progressively better instruments for measuring an invariant quantity called "value".
The theoretical object itself changes with the analytical system.
Classical economics asks one set of questions about production, surplus and distribution.
Marx reconstructs those relationships within another theoretical system.
Marginalism reformulates value through scarcity, preferences and marginal substitution.
Later debates over capital and distribution again challenge the conceptual foundations of the theory.
There is therefore no theory-independent instrument that reveals:
\[ \boxed{\text{Economic Value}=€X} \]
before economic theory enters the analysis.
Economic value is not merely difficult to observe.
Its measurement requires a prior theoretical definition of what is to be measured.
One obvious response is to define economic value as market value.
Then
\[ E=P. \]
This is useful in many contexts.
But it does not solve the general problem.
First, many economic resources have no independent market price.
An integrated production facility may have value as part of a going concern without possessing a continuously observable standalone market price.
Internally generated software, organizational capital, customer relationships, specialized knowledge and human capital provide even clearer examples.
Second, even where markets exist, price depends on circumstances.
A transaction price may depend upon:
\[ \text{liquidity}, \quad \text{information}, \quad \text{bargaining power}, \quad \text{market structure}, \quad \text{time}, \quad \text{expectations}. \]
Third, defining economic value as market price risks making the benchmark tautological.
If
\[ E=P, \]
then asking whether market price correctly measures economic value becomes impossible by definition.
Price is no longer evidence about value.
It is value.
That is a legitimate definition for some purposes, but it is not a neutral solution to the historical problem of value.
Modern finance offers another candidate:
\[ E = \sum_{t=1}^{T} \frac{\mathbb{E}(CF_t)} {(1+r_t)^t}. \]
This is enormously useful.
But it should not be confused with direct observation.
To calculate \(E\), one must specify
\[ \mathbb{E}(CF_t), \]
the expected future cash flows, and
\[ r_t, \]
the appropriate discount rates.
Neither is generally observable for the complete future life of an economic resource.
Consequently,
\[ E=f(\text{expectations},\text{model},\text{discount rate}). \]
Two rational analysts may use different expectations or models and obtain
\[ E_1\neq E_2. \]
Neither difference necessarily represents an arithmetic error.
The economic value is model-dependent.
Present-value techniques therefore provide valuations, not direct observations of value.
This creates an important refinement of the Accounting Bias framework.
The theoretical definition remains
\[ AB=A-E. \]
But \(A\) and \(E\) have fundamentally different epistemic properties.
Accounting representation is usually observable:
\[ A=\text{observed}. \]
Economic value is generally latent:
\[ E=\text{unobserved}. \]
Therefore,
\[ AB=A-E \]
is theoretically meaningful without necessarily being directly measurable.
This distinction is crucial.
It prevents the framework from making the unjustified claim that the researcher knows the "correct" economic number and merely needs to compare accounting against it.
Instead, Accounting Bias identifies a measurement relationship.
Accounting transforms economic phenomena through recognition, measurement, timing, allocation and estimation rules:
\[ A=\mathcal{A}(E,\mathcal{R},\mathcal{M},\mathcal{T},\mathcal{S}). \]
The research question becomes:
How does the accounting system transform economically relevant phenomena into reported numbers?
That question remains meaningful even when \(E\) cannot be observed directly.
This distinction is particularly important.
If economic value cannot generally be observed directly, it does not follow that economic value is meaningless.
The relevant distinction is between
\[ \text{conceptual definition} \]
and
\[ \text{empirical identification}. \]
Economic theory can define an economic benchmark without guaranteeing that the benchmark is directly observable.
Accounting Bias can therefore exist theoretically even when its exact cardinal magnitude cannot be recovered empirically.
Suppose internally generated knowledge contributes substantially to future production but fails the relevant recognition conditions.
The researcher may have strong theoretical reasons to believe that accounting representation omits an economically relevant resource:
\[
A
But this does not imply that the researcher knows whether
\[
E-A=€10\text{ million}
\]
or
\[
E-A=€100\text{ million}.
\]
Direction may sometimes be theoretically defensible when magnitude is not.
This distinction between existence, direction and
magnitude is fundamental.
The empirical framework developed in Accounting Bias follows directly
from this limitation.
If \(E\) is unobservable, empirical research should not manufacture an
apparently precise estimate of
\[
A-E
\]
without acknowledging the additional valuation model required to construct
\(E\).
Instead, observable variables can identify exposure to accounting mechanisms
capable of generating divergence.
For example,
\[
\frac{R\&D}{TA}
\]
does not measure the value of unrecognized knowledge.
It measures exposure to a recognition mechanism.
Similarly,
\[
\frac{PPE}{TA}
\]
does not measure the difference between the economic value and carrying amount
of productive assets.
It measures exposure to historical-cost measurement, depreciation and
impairment mechanisms.
Accrual intensity captures exposure to timing and allocation.
Goodwill intensity captures exposure to acquisition accounting and impairment
estimation.
Deferred-tax positions capture another form of recognition and timing
interaction.
The empirical architecture consequently becomes:
\[
\boxed{
\text{Theory of economic reality}
\rightarrow
\text{Accounting rule}
\rightarrow
\text{Potential divergence}
\rightarrow
\text{Observable exposure}
\rightarrow
\text{Empirical implication}
}
\]
This is deliberately weaker than claiming direct observation of economic value.
It is also more defensible.
The non-measurability of economic value leads to another important conclusion.
Accounting Bias should not be interpreted as accounting error.
If \(E\) were objectively observable and accounting reported some different
quantity \(A\), then one might reasonably ask why accountants did not simply
report \(E\).
But that is not the world in which accounting operates.
Accounting must construct operational measurements despite uncertainty about
economic value.
It therefore requires rules.
Recognition thresholds determine what enters the accounts.
Measurement bases determine how recognized items are quantified.
Depreciation and amortization allocate amounts across time.
Impairment rules determine when carrying amounts are revised.
Provisions transform uncertain future obligations into current accounting
quantities.
These mechanisms produce
\[
A.
\]
Their purpose is not necessarily to discover a metaphysical "true value".
They construct a standardized, auditable and institutionally usable
representation of economic activity.
Accounting Bias arises because every such representation necessarily selects a
measurement architecture.
This suggests a broader interpretation.
Accounting can itself be understood as a theory of economic representation.
Consider the balance sheet:
\[
Assets
=
Liabilities
+
Equity.
\]
The identity is exact within the accounting system.
But the quantities entering the identity depend upon prior decisions concerning:
\[
\text{recognition},
\quad
\text{measurement},
\quad
\text{classification},
\quad
\text{timing},
\quad
\text{estimation}.
\]
The accounting equation is therefore mathematically exact while its relationship
to economic reality remains mediated by accounting rules.
There is no contradiction here.
Accounting can be internally exact and economically incomplete simultaneously.
Indeed, that distinction is at the heart of Accounting Bias.
Dobb's history of value theory therefore has a consequence extending beyond the
history of economic thought.
If Smith, Ricardo, Marx, Jevons and subsequent economists disagree about the
determinants and conceptual structure of value, then financial reporting cannot
simply import a universally accepted economic value from economics.
Economics gives accounting multiple possible benchmarks:
\[
E_{\text{production}},
\]
\[
E_{\text{market}},
\]
\[
E_{\text{utility}},
\]
\[
E_{\text{discounted cash flow}},
\]
\[
E_{\text{replacement}},
\]
and potentially others.
These need not coincide:
\[
E_1\neq E_2\neq E_3.
\]
Choosing among them is itself a theoretical and institutional decision.
Accounting standards therefore do something unavoidable.
They select measurement conventions from a world in which economic valuation is
underdetermined.
Historical cost, fair value, amortized cost, value in use and other measurement
bases can be interpreted as institutional responses to different aspects of
this underlying problem.
Accounting is consequently not merely a defective version of economic
valuation.
It solves a different problem.
Economics asks:
\[
\textit{What determines economic value?}
\]
Accounting must ask:
\[
\textit{What quantity can be recognized, measured, verified and reported?}
\]
Those questions overlap.
They are not identical.
The relationship can finally be expressed as follows.
Suppose economic reality is represented by
\[
E.
\]
Economic theory does not give us direct access to \(E\).
Instead, competing theoretical frameworks produce valuation mappings:
\[
V_1(E),V_2(E),\ldots,V_n(E).
\]
Accounting introduces another mapping:
\[
A=\mathcal{A}(E).
\]
The empirical researcher observes some consequences of these mappings, but not
necessarily \(E\) itself.
The problem is therefore not simply:
\[
\text{Does }A=E?
\]
It is:
\[
\boxed{
\text{How does } \mathcal{A}
\text{ transform economic reality into an observable representation?}
}
\]
Accounting Bias is the conceptual device used to study that transformation.
The inability to observe \(E\) does not invalidate the framework.
It determines the limits of what the framework can claim empirically.
The history of economic thought provides an important warning for any theory of
accounting measurement.
There is no simple, theory-independent and universally observable quantity
called economic value waiting behind every accounting number.
Classical political economy sought the determinants of value primarily through
production and distribution.
Marx reconstructed the relationship between value and observable prices.
Marginalism shifted the analysis toward scarcity, preferences and marginal
choice.
Modern valuation approaches often define value through expected future cash
flows and discount rates.
Each framework illuminates something important.
None provides a universal instrument with which economic value can simply be
observed.
For the Accounting Bias framework, the implication is fundamental:
\[
\boxed{
AB=A-E
}
\]
should be understood first as a theoretical relation, not as a
promise that \(E\) is directly measurable.
Accounting representation \(A\) is observable.
Economic value \(E\) is generally latent, model-dependent, and sometimes
conceptually dependent upon the theory of value being employed.
The task of empirical research is therefore not to manufacture a false precision
around \(E\).
It is to identify the accounting mechanisms through which economic phenomena
become reported numbers and to study the observable consequences of those
mechanisms.
This leads to a more modest but more powerful interpretation of Accounting Bias.
Accounting does not fail because it cannot perfectly measure economic value.
Accounting exists partly because economic value cannot be perfectly measured.
Dobb, Maurice.
Theories of Value and Distribution since Adam Smith: Ideology and Economic Theory.
Cambridge University Press, 1973.
11. From Measurement to Exposure
12. Accounting Is Not Wrong Because It Does Not Measure \(E\)
13. Accounting as One Theory of Representation
14. From Dobb to Accounting Measurement
15. The Fundamental Measurement Problem
Conclusion
Reference