The implications of absorption cost accounting and production decisions for firms’
future performance and valuation
Chandra Seethamraju **
We appreciate the helpful comments of seminar participants at New York University, University of
Arkansas and participants at the accounting department brown bag at Washington University.
** Corresponding Author. #1 Olympian Way, Campus Box 1133, Washington University, Saint
Louis, MO 63130.
In this paper, we evaluate a comprehensive set of inventory related signals, the association
between those signals and future firm performance and how the stock market responds to these
signals. There is mixed evidence in the literature on how the market perceives and reacts to
inventory build-ups and possible income manipulation. Lev and Thiagarajan (1993) conclude that
the stock market views their measure of excess inventory growth as a negative signal, while
Jiambalvo et al. (1997) find evidence that stock returns are positively associated with their measure
of overproduction. In our view, the stock market perception of overproduction should also be
affected by the materiality of the fixed costs that get absorbed into ending inventory. We provide
evidence that future accounting performance (as measured by ROA) of high relative fixed cost/high
overproduction firms is significantly negatively associated with overproduction, while this
relationship is positive for other firms. We find that the stock market reacts in a rational manner
and that there is a substantial reduction in the positive valuation impact of overproduction for firms
with high relative fixed costs and high overproduction.
We also examine the impact of order backlog, as also its interaction with overproduction
and high relative fixed cost structure. We find that, in general, for firms that have a “bad news”
order backlog signal, overproduction is more negatively associated with future performance (as
measured by ROA) relative to firms that did not have such a signal. We also report strong
evidence that the market reacts rationally and discounts reported earnings of firms with ‘bad
news’ order backlog signals, relative to other firms.
In this paper, we evaluate a comprehensive set of inventory related signals, the association
between those signals and future firm performance and how the stock market responds to these
signals. Most management accounting texts discuss the fact that manufacturing firms can increase
reported income by producing goods in excess of the quantity required to meet current demand
(see, for e.g. Horngren, Datar and Foster, 2002, p.287). One reason for such potentially sub-
optimal behavior is that if managers are compensated based on reported operating income, they
may be tempted to over-produce in order to increase reported operating income. Other reasons for
firms to indulge in this type of opportunistic behavior could include stock market pressures to
meet analysts’ forecasts of earnings. Roychowdhury (2004) investigates whether firms succeed in
avoiding losses by engaging in managing sales via, perhaps, relaxing credit terms and offering
sales discounts, reduction of discretionary expenditures, such as advertising and R & D, and
inventory over-production. He reports that inventory over-production is positively associated with
the avoidance of losses and this effect is more pronounced among manufacturing firms.
Since most firms use absorption costing (as opposed to variable costing), such
overproduction results in allocation of part of current period fixed manufacturing overhead to
ending inventory rather than to cost of goods sold. This in turn reduces cost of goods sold and
increases reported income for the period. Therefore, the component of earnings attributable to
over-production may reflect potential opportunistic behavior and may be viewed as being of
lower quality than the remaining component of earnings. There is mixed evidence in the literature
on how the market perceives and reacts to inventory build-ups and possible income manipulation.
Lev and Thiagarajan (1993) examine the value-relevance of multiple signals of future firm
performance identified in Value Line analyst reports. One of the signals of interest to them was
the excess of annual percentage inventory growth over annual percentage sales growth. Lev and
Thiagarajan conclude that the stock market views excess inventory growth as a negative signal,
indicating possible production for inventory and/or problems with inventory obsolescence and
turnover. Abarbanell and Bushee (1997) confirm that excess inventory growth had a negative
association with future earnings.
It is also possible that managers may produce more than current period demand in order to
be able to meet future anticipated demand. In such an instance, overproduction would relate to
future firm performance and would not reflect opportunistic behavior. Jiambalvo, Noreen and
Shevlin (1997) (henceforth referred to as JNS) explore this issue using a large sample of
manufacturing firms and find evidence that stock returns are positively associated with their
measure of overproduction, CPAI (change in percent of production added to inventory)1. In their
view, this finding is consistent with stock market participants viewing CPAI as a leading positive
indicator of firm performance. JNS only examine the quantity added to inventory. The literature,
however, has not examined the impact of the interaction of the components of overproduction
which impact the income statement– the quantity added to inventory and the relative fixed costs
level which impacts the dollar amount absorbed in ending inventory.
In this paper, we provide one of the first sets of empirical results that show that the
arguments of both Lev and Thiagarajan (1993) and JNS (1997) can coexist. In our view, the stock
markets perception of CPAI should also be affected by the amount of fixed costs that get
absorbed into ending inventory. In other words, if a firm had a high fixed cost structure and
produced in excess of current demand, the impact on income for such a firm would be much
greater than for a firm which did not have a high fixed cost structure. In this study, we propose a
1 JNS also control for the Lev and Thiagarajan (1993) measure of abnormal inventory growth. They show that both
signals are incremental to each other.
parsimonious proxy (called, Fixed Asset Intensity) for a firms’ fixed cost structure, measured as
the ratio of a firms’ gross fixed assets to total assets in a given year. We identify high relative
fixed cost firms as those firms with Fixed Asset Intensity greater than the median Fixed Asset
Intensity for all firms in the firms’ 2 digit SIC code in that year. We identify high over-production
firms as those firms with CPAI greater than the median CPAI for all firms in the firms’ 2 digit
SIC code in that year. We provide evidence that future accounting performance (as measured by
ROA) of high relative fixed cost/high overproduction firms is significantly negatively associated
with overproduction, while this relationship is positive for other firms. More importantly, to the
extent that such firms add to inventory opportunistically, they could be masking
performance/operating problems by indulging in this behavior in the current period. It would not
be a sustainable strategy for firms to overproduce over multiple periods when the overproduction
is not matched by future revenues. Therefore, the positive impact of overproduction on
performance (for firms producing opportunistically) in the current period should be reversed in
We hypothesize that the stock market discounts the positive impact of CPAI found in JNS
for firms with high relative fixed costs (and high overproduction) compared to other firms. The
reason for this is that this component of earnings should be of lower ‘quality’, given that for these
firms a material amount of current fixed overhead has been absorbed into ending inventory rather
than cost of goods sold. We find that the stock market reacts in a rational manner and report
strong evidence in support of our hypothesis. There is a substantial reduction in the positive
valuation impact of CPAI for firms with high relative fixed costs and high overproduction. The
market seems to make adjustments for the lower ‘quality’ of reported earnings.
The stock markets perception of overproduction (which is essentially an inventory build-
up) could also be impacted by signals related to future sales expectations. Order Backlog is an
additional inventory related disclosure that potentially captures future sales expectation, since it is
the dollar value of unfilled sales orders by the firm at the end of the period.2 Rajgopal et al.
(2003) report that order backlog is value-relevant and has a positive association with future
earnings. We examine the impact of order backlog, as also its interaction with CPAI (another
leading indicator of future performance, as per JNS) and high relative fixed asset intensity.
We find that, in general, for firms that have a “bad news” order backlog signal, CPAI is
significantly more negatively associated with future performance (as measured by ROA) relative
to other firms. In addition, we provide strong evidence that the market reacts rationally and
discounts reported earnings of firms with ‘bad news’ order backlog signals, relative to other
Levitt (1998) voiced concerns that the motivation for firms to meet Wall Street earnings
expectations could be overriding common sense business practices. This paper falls into the
stream of literature that examines income manipulation via ‘real decisions’, i.e. when
management intervenes in the “normal” earnings process via potentially sub-optimal managerial
decisions, such as excessive production, scaling back research and development expenditures
(Bushee (1998), Dechow and Sloan (1991)), or timing of asset sales (Bartov, 1993).
Our paper contributes to the literature in three ways. First, our study fills a gap in the
literature by explicitly examining whether the market understands the implications of absorption
costing and that inventory over-production could have subsequent period performance effects. In
addition, un-justified over-production could be evidence of short-sighted decisions by managers,
and understanding whether the market identifies such sub-optimal decisions is an important
2 Firms are required to disclose Order Backlog when the number is deemed to be material.
question. Second, we examine an important additional dimension of inventory overproduction
which was, hitherto, not explicitly examined by the literature – relative fixed cost structure of
firms and its interaction with inventory build-ups in impacting reported net income. Finally, we
examine a comprehensive set of inventory related variables and study them in conjunction with
each other, and not in isolation, as the literature has tended to do so. This enables us to develop a
richer understanding of how the stock market perceives the various inventory related signals and
their interactions with each other.
The rest of the paper is organized as follows: in Section 2, we provide the motivation and
develop the hypotheses. In Section 3, we describe our sample, and in Section 4, we describe our
findings for the main sample. In section 5, we describe our findings for the order backlog sample.
Conclusions are provided in Section 6.
2. Motivation and Hypotheses development
The issue of absorption costing vs. variable costing methods has not received substantial
attention in the empirical financial accounting literature. Instead, the research has primarily
focused on the issues of choice of valuation models for inventories (LIFO vs. FIFO, see, for
example Kinney and Wempe (2004), Hunt et al. (2000), Pincus (1997), among more recent
Lev and Thiagarajan (1993) examine various fundamental signals of future performance,
such as excess receivables growth over sales growth, excess inventory growth over sales growth,
order backlog, capital expenditures, gross margin, etc., in relation to the contemporaneous
returns. Their motivation for using these signals was based on analyst reports from Value Line
analysts, wherein these fundamental signals were discussed. According to Lev and Thiagarajan
(1993), these analysts view excess percentage change in inventory over percentage change in
contemporaneous sales as a negative signal, reflecting an unjustified or unsustainable build-up of
inventories. Lev and Thiagarajan (1993) find that excess inventory build-ups are negatively
associated with current returns, consistent with analysts’ assessment of this signal. Abarbanell
and Bushee (1997) report that inventory build-ups are negatively associated with future
performance. Roychowdhry’s (2004) measure of inventory over-production is closely related to
the Lev and Thiagrajan (1993) abnormal inventory growth measure. He uses the Dechow et al.
(1998) accrual model to derive the expected cost of goods sold and expected annual inventory
change as a function of the firms’ prior period sales and sales change. Roychowdhury’s (2004)
measure of abnormal production explicitly incorporates the cost of and prior sales history in the
production costs expectation model. It does not, however, explicitly model how fixed costs are
factored into inventory valuation.
Another study that has attempted to more explicitly tie variations in fixed cost structure to
inventory valuation is JNS (1997). They provide an analytical model of the differential between
absorption and variable costing earnings and develop an explicit measure of the effect of
“production for inventory” on this differential. They call this measure Change in Percentage of
Production Added to Inventory (CPAI).3 JNS show that CPAI is positively associated with
contemporaneous CARs, incremental to the annual earnings surprise. Moreover, they show that
CPAI is informative incremental to the Lev and Thiagarajan (1993) excess inventory growth
measure, suggesting that these two signals are independent of each other. JNS interpret their
results as indicating that CPAI is viewed by the market as providing favorable information about
future performance, incremental to the unexpected earnings and past sales information. JNS also
3 For the derivation of CPAI refer to the Appendix section of JNS (pages 91-93).
attempt to allow fixed manufacturing overhead to vary across firms by running industry-based
regressions and controlling for firm size. They do not find a differential impact of CPAI in these
The findings of Lev and Thigarajan (1993) and Abarbanell and Bushee (1997) on one
hand, and JNS (1997) on the other hand, seem to indicate that the two different measures of
overproduction examined in these studies are viewed differently by the market. The former two
studies report that their measure is viewed as a negative signal, while JNS report that their
measure is viewed as a positive signal by the market. In addition, JNS suggest that these two
signals appear to be incremental to each other. The variable used in Lev and Thiagarajan (1993)
compares changes in inventory to contemporaneous changes in sales. However, there could be
changes in inventory due to anticipated changes in sales in future periods and their variable does
not capture that effect. None of these studies controls explicitly for the absorption of fixed
manufacturing overhead (FMO). We argue that if FMO is brought into the analysis, situations
where these variables could provide different signals can be studied. Since the two variables (Lev
and Thiagarajan (1993) and JNS (1997)) are measured differently, we further study the variable
used in JNS (namely, CPAI) and examine under what conditions CPAI is viewed positively and
when it is viewed negatively by the market.
2.2. Impact of Firms’ Fixed manufacturing overhead
For a manufacturing firm, the biggest reason for fixed manufacturing overhead is its own
production and manufacturing facilities. The higher the firm’s investment in fixed assets
associated with inventory production, the higher is the amount of expense associated with these
fixed assets and is considered fixed manufacturing overhead allocated to inventory. Such expense
includes depreciation charges, repair and maintenance expenses on the manufacturing facilities,
and the like. These charges are not separately disclosed in the financial statements as they are
included in the cost of goods sold and in the ending inventory balance.4 The JNS model indicates
that CPAI is one of the two components impacting the differential between absorption costing
earnings and variable costing earnings. The second component is firms’ fixed manufacturing
overhead. JNS explicitly assume that fixed manufacturing overhead (as a percentage of the
deflator, market value of equity) of all firms’ is a cross-sectional and time-series constant. They
attempt to control for fixed costs by re-estimating their analysis across industries (2 digit SIC
code level) and size deciles, and their results are unchanged in that analysis. JNS acknowledge
that their approach has limitations since it does not explicitly incorporate firms’ fixed cost
properties in the analysis.
We feel that fixed manufacturing overhead (FMO) is an important factor that needs to be
integrated into the analysis, while being cognizant of the fact that this number is not available to
outside users of financial statements. Hence, we model FMO’s cross-sectional variation in our
tests in the following manner. We use the ratio of gross fixed assets to total assets of the firm in
year t, as a measure of a firm’s fixed cost structure and call this ratio Fixed Asset Intensity
(henceforth referred to as FAI). We use gross fixed assets since gross fixed assets are free of
allocation rules used for depreciation and provide for a more homogenous way to compare this
variable across firms. It is also not clear which number investors are focusing on in order to
determine capital investment and its impact on fixed costs5. Thus, FAI is a proxy for the
magnitude of FMO allocated to inventories in the end of the period.
4 Recently, as part of its International Convergence Project, FASB issued Statement 151 that becomes effective for
companies with year end after June 15, 2005, that modifies full absorption costing accounting to exclude certain
abnormal amounts of the idle facility expense, freight, handling costs and wasted materials to be expensed in the
5 In addition, we also use the ratio of net fixed assets to total assets as an alternative measure of Fixed asset intensity.
We redo the main tests reported in the paper using this variable and the results are essentially unchanged.