Fairfield et al. (2003) 認為公司當年度淨營運㈾產增加，將會使次年度的 ㈾產報酬率㆘降，其原因為（1）邊際報酬遞減以及（2）保守原則。該論文也 發現股市對㆘列兩㊠㈶務㈾訊均過度反應：應計㊠目與長期淨營業㈾產的成 長。 本論文則認為若公司投㈾決策比較合乎理性投㈾的假設，則公司新計畫的預期 投㈾報酬率應該會高於其㈾㈮成本率，所以平均而言，今年度㈾產報酬率低於 ㈾㈮成本率的公司若增加淨營業㈾產，則次年度的㈾產報酬率應該會㆖升；反 之，今年度㈾產報酬率高於㈾㈮成本率的公司若增加淨營業㈾產，則次年度的 ㈾產報酬率是否增加就無法定論，因為新投㈾的預期報酬率不㆒定比舊投㈾的 報酬率高。另本文亦討論 FWY 所提之穩健會計論點：該文認為公司增加營運 ㈾產，次年績效將會降低是因為穩健會計，但本文認為穩健會計將使投㈾計畫 初期無法㈾本化支出較多，因此未來年度相對績效反而將增加，這似乎也與 FWY 之預期相反。本文之實證結果支持我們的兩㊠不同論點。

Fairfield et al. (2003) propose that the reasons why firms investing more in net operating assets at current year will experience lower one-year ahead ROA arise from both diminishing marginal returns on investment (Stigler, 1963) and conservative accounting (Penman, 2001). They also reveal that investors appear to equivalently overprice accruals and growth in long-term net operating assets. Their evidence shows that firms with a growth in long-term net operating assets would be, on average, less profitable in the following year. However, we argue that rational firms should invest in long-term net operating assets when the internal rates of returns of the new projects are higher than the required rate of returns. For firms with current ROA lower than the required rate of returns, their future ROA should be, on average, increasing if they increase investments. For firms with current ROA higher than the required rate of returns, the future ROA could be increasing or decreasing, since the result depends on the relative levels of the current ROA and the internal rate of returns of the new projects. Empirical evidence is more consistent with our rational investment argument. The evidence is also inconsistent with FWY’s argument regarding conservative accounting bias. When firms increases operating asset, conservative accounting bias means that expenditures should be expensed when incurred, not to defer them into the future. This could result in a higher one-year-ahead ROA, unless these expenditures persist into the next year. Since most projects involve some form of start-up costs, we believe that, under possible conservative accounting, the accounting treatment for the second year of the asset life would be less conservative than that of the first year. So, the second year ROA will not be, on average, decreasing as FWY expected.

If firms’ investing decisions are rational, then the purpose of increasing net operating assets is certainly to enhance firm performance. Accordingly, we should exptect a better outlook for firms’ returns on asset (ROA). However, Fairfield et al. (2003) (FWY, hereafter) found that an increase in net operating assets would result in a decrease in firms’ ROA. They argue that diminishing marginal returns on investments arise when firms exploit their most profitable investment opportunities before undertaking less profitable investments. Alternatively, increasing marginal returns in divestment arise when firm divest their least profitable investments. In addition, they also show that a conservative bias in accounting procedure results in investments that appear relatively less profitable in early years and more profitable in later years, biasing accounting rates of return on new investments downward relative to the returns in existing investments. Thus, they conclude that when firms invest more in long-term net operating assets, both conservative accounting bias and diminishing marginal returns to operating assets would, on average, reduce near-term profitability. In this paper, we argue that diminishing return on assets does not fully describe firms’ investing behavior. Rational firms should invest in long-term net operating assets when the internal rates of returns of the new projects are higher than the required rate of returns. For firms with current lower than the required rate of returns, the future should be, on average, increasing if they increase investments. For firms with current higher than the required rate of returns, the future could be increasing or decreasing, since the result depends on the relative levels of the current and the internal rate of returns of the new projects. That is, if the internal rate of returns of the new projects is higher than the current , then the future could be increasing. In contrast, if the internal rate of returns of the new projects is less than the current , the future could be decreasing. FWY contend that their evidence is also consistent with the conservative accounting bias. Nonetheless, conservative accounting bias only states that costs with uncertain future economic benefit and some indirect costs related to the acquisition of long-term assets tend to be expensed when incurred. Expenditures which are expensed when incurred could result in a higher one-year-ahead unless these expenditures persist into the next year. This seems to contradict FWY’s argument. For the accounting treatment of the capitalized direct costs, it would be equally conservative for the first year and the second year of the asset life if straight-line method is used for recording the depreciation expenses. The accounting treatment of the first year of the asset life is actually more conservative then that of second year of the asset life when accelerated methods are implemented. Therefore, we believe that, under possible conservative accounting, the accounting treatment for the second year of the asset life would be less conservative than that of the first year. So, the second year ROA will be, on average, increasing as compared to the first year. Market participants’ reaction to certain financial information is of interest to not only the market regulators but also the market participants themselves. Our results show that the stock market is potentially inefficient when reacting to the change in net operating assets.