Recognition Versus Disclosure in The Oil and Gas Industry

For FC firms the market generally observes the write-down at the write-downs announcement date and for SE firms the market can estimate the disclosed write-down at the 10-K filing.1 When

Recognition Versus Disclosure in The Oil and Gas Industry

David Aboody

Anderson Graduate School of ManagementUniversity of California, Los AngelesLos Angeles, CA 90095-1481E-mail: david.aboody@anderson.ucla.edu

July 1996

I wish to thank Shlomo Benartzi, Nicholas Dopuch, Stephen Hansen, Robert

Holthausen, Pat Hughes, Gil Mehrez, Judy Rayburn, and participants in the JAR

conference for their helpful comments I also wish to thank the anonymous referees and editor for their patience and helpful comments I am especially grateful to Keith Klaver from Price Waterhouse LLP for providing me with a practitioner’s view of oil and gas accounting

Recognition Versus Disclosure in The Oil and Gas Industry

1 Introduction

This paper investigates whether recognition and disclosure are equivalent

in their pricing consequences in the oil and gas industry The investigation concentrates on the oil and gas industry because Securities and Exchange Commission (SEC) regulation SX 4-10 provides a unique opportunity for testing the stock price consequences of recognition versus disclosure The SEC

regulation requires the firm-specific effect of a macroeconomic event to be formally recognized in the financial statements for oil and gas firms adopting the full cost method and disclosed in footnotes for firms following the successfulefforts method Results of multivariate tests indicate that in an investment setting footnote disclosure is not equivalent to recognition

Whether users of financial statements distinguish between recognition and disclosure is a central question in the debate over accounting standards However this question was never explicitly empirically tested (for a literature review, see Bernard and Schipper [1994]) Existing empirical studies (e.g., Amir[1993], Barth [1994]) investigate the association between the firm’s market value and an estimate of the disclosed item However, although the coefficient estimates on the disclosed item have the predicted sign, the coefficient

estimates are significantly different (in both directions) from the theoretical predictions The results may be attributed to market inefficiency, measurementerror in the disclosed items, or because disclosed items are less reliable than recognized items The experimental literature (Harper, Mister, and Strawser

[1987,1991]) suggests that whether an item is recognized or disclosed

influences financial statements users’ perceptions However, the experimental studies do not address the issue of whether different users’ perceptions will result in different pricing in the marketplace

The accounting treatment for oil and gas firms provides a unique

opportunity for testing the pricing implications of recognition and disclosure SEC regulation SX 4-10 allows firms adopting the FC method to capitalize all costs associated with property acquisition, exploration, and developments

activities However, if the net capitalized cost of FC firms exceeds the net

discounted future cash flows from proved oil and gas reserves (termed

“ceiling”), the excess is an ordinary loss The SEC regulation requires SE firms

to disclose in footnotes the “ceiling” and their net capitalized asset value However, the SEC and Generally Accepted Accounting Principles (GAAP) force

SE firms to recognize a write-down only if the capitalized costs exceed the net

undiscounted future cash flows from proved oil and gas reserves

Consequently, if the net capitalized costs exceed the “ceiling” but are less than the undiscounted cash flows, an FC firm must write-down its assets to the

discounted cash flow while an SE firm will only report an as-if write-down in its footnotes

This study focuses on a sample of 21 FC firms that formally recognize a write-down and a sample of 50 SE firms that discloses the economic loss in theirfootnotes I conduct an event study investigating the cross-sectional variation

of stock returns at the write-downs announcement date and the 10-K filing date

For FC firms the market generally observes the write-down at the write-downs announcement date and for SE firms the market can estimate the disclosed write-down at the 10-K filing.1 When investigating the market reaction, I test whether the market responds similarly to FC firms recognizing a loss and SE firms disclosing one.2

Pooled cross-sectional regression results document, at the write-downs announcement date, a significant negative market reaction to firms recognizing

a write-down At the 10-K filing date there is no significant market reaction to firms disclosing a write-down In addition, at the 10-K filing date the increase in the probability of debt covenant violation is negatively associated with stock returns for both SE and FC firms The main result that investors price differentlyrecognized and disclosed write-downs is robust to several competing

hypotheses

My results suggest that information disclosed in footnotes to the financial statements is used differently by investors from information recognized in the primary financial statements The main finding is that the market reaction to firms recognizing a write-down significantly differs from the reaction to firms disclosing it One explanation is that the FC firms’ write-downs are visible while favorable information offsets SE firms’ write-downs released in the 10-K An alternative explanation, given the material impact of write-downs on firms’ net

1 All but two firms disclosed the write-down at the earnings announcement date The 3-day return for the 2 firms disclosing the write-down before the earnings release was -66.67% and -12.5%.

2 In a Wall Street Journal article dated April 17, 1992 Catherine Montgomery, an analyst with Donaldson, Lufkin & Jenrette, discusses

the market reaction to FC firms recognizing a loss "I think that serious investors, institutions and analysts understand these downs and what they say about the companies forced to take them But the average investor out there has a knee-jerk response and stock prices may be affected."

write-equity, is that investors concentrate on firms’ net equity for valuation purposes.3

The remainder of the study is organized as follows Section 2 includes therelated oil and gas accounting background and the method for estimating the

SE firms’ economic loss Empirical tests are developed in section 3 Section 4 details the sample selection and descriptive statistics Results and diagnostic tests are presented in section 5, and summary and concluding remarks are presented in section 6

3 In an interview with Mr Braverman, an investment officer with Standard & Poors, the Wall Street Journal (December 21, 1995) reports the following: “Nor says Mr Braverman should they [investors] ignore that a big write-off knocks down a company’s book value, known as shareholders’ equity, an important yardstick

of a company’s worth.”

2 Accounting Background and the Disclosed Write-Down for SE Firms 2.1 Accounting background Securities and Exchange Commission (SEC)

regulation SX 4-10 prescribes financial reporting for firms engaged in oil and gasproducing activities According to SX 4-10, firms adopting the FC method (from hereon FC firms) may capitalize all costs associated with property acquisition, exploration and development activities Firms are required to capitalize the costs within the appropriate cost center that is based on a country-by-country basis In contrast, firms following the successful efforts method (from hereon SEfirms) are allowed to capitalize the same costs only if they result in an increase

of proved oil and gas reserves

The regulation states that FC firms must perform a “ceiling test” at the end of each quarter The ceiling for each country is the sum of: (1) the present value of proved oil and gas revenues from estimated production of proved oil and gas reserves, based on the oil and gas prices at the end of each quarter and a discount rate of 10%, (2) the cost of properties not being amortized, and (3) the lower of cost or estimated fair value of unproved properties included in the costs being amortized, less (4) income tax effects related to the difference between book and tax basis of the properties involved.4

Subsequently, if the net capitalized costs within a cost center, less relateddeferred income taxes, exceed the cost center ceiling, the excess is an ordinary loss FC firms cannot reinstate the loss (termed a write-down) due to a

subsequent increase in the cost center ceiling The SEC did not apply this

4 Cited from regulation SX 4-10, Full Cost Method paragraph (4).

5

recognition requirement to SE firms because it contended that the write-down event should be rare Therefore, SE firms recognize a write-down only if their net capitalized costs are higher than the net undiscounted value of their proved oil and gas reserves.5 Consequently, the different reporting rule causes FC firms

to recognize in the income statement the excess of the net capitalized cost overthe ceiling while SE firms need only disclose the capitalized cost and ceiling in the supplementary unaudited part of the financial statements

2.2 Calculation of "as-if" write-downs for successful efforts firms

Regulation SX 4-10 requires FC firms to recognize write-downs while for SE firmsinvestors can only infer the economic loss from footnote disclosure This

section details how footnote information for SE firms is used to calculate their economic loss, and the appendix details a specific example of how the footnote disclosure to calculate the “as-if” write-down is used

As previously discussed four components compose the ceiling The

supplementary unaudited section of the financial statements provides

information on two components: the present value of future net revenues from the oil and gas proved reserves and the income tax effect related to them The audited part of the financial statements provide the other two components, cost

of unproved properties and properties not being amortized

5 Before December 15, 1995 managers of SE firms with net capitalized assets higher than the undiscounted value of proved reserves had discretion over whether the assets should be written down to the discounted

or undiscounted value I found only one SE firm stating that if a write-down occurs, the asset value will be

reduced to the discounted value After December 15, 1995, FAS 121 (Accounting for the Impairment of Long-lived Assets) requires the asset to be written down to its discounted value.

6

The as-if write-down is the excess of the disclosed ceiling amount over thenet after-tax capitalized cost of proved oil and gas assets The audited part of the financial statements reports the before-tax net capitalized cost of oil and gas properties However, the sample firms do not detail the amount of deferredincome taxes associated with the capitalized costs in their tax footnote

Therefore the pretax as-if write-down is calculated by adding back to the ceiling its fourth component, namely the present value of income taxes The sample of

50 SE firms includes 16 firms that explicitly disclose the present value of incometaxes and 34 firms that disclose the future value of income taxes For the 34 firms, the discount rate they use to discount their future net cash flows is

applied to the future value of income taxes Since, this procedure can

underestimate or overestimate the write-down given the firm’s specific pattern

of cash flows, the calculation is repeated by using the minimum of the discount rate, over the past three years resulting in the elimination of two as-if write-downs Finally, out of seventeen firms operating outside the United States, seven firms did not provide a country-by-country breakdown of the information Therefore, because the as-if write-down is estimated on an aggregate level the write-down is underestimated for these firms

In this study, the pretax economic loss for SE firms is the excess of the pretax net capitalized costs over the pretax ceiling Consequently, users of SE firms’ financial statements could also perform the above calculation and arrive

at an “as-if” write-down for the SE firms’ properties Therefore, SE firms are considered the disclosing firms, and FC firms are the recognizing firms

7

3 Research Design for Testing Recognition and Disclosure in an

Investment Setting

In an investment setting, recognition and disclosure could have different implications for valuation of stocks for several reasons This section develops the reasons and the research design applied to address them

3.1 Reliability of recognized and disclosed items Statement of

Financial Accounting Concepts No 5 (SFAC 5) specifies four criteria for

recognition Among others, the item must be relevant, in the sense of affecting decisions, and measured with “sufficient reliability” to be recognized

Therefore, the act of recognizing an item is revealing with respect to the

underlying relevance and reliability of the information

A major component in the ceiling calculation for firms in the oil and gas industry is the quantity of proved oil and gas reserves An independent

consulting firm calculates and provides the quantity at the end of each year The amount of the write-down also depends on the book value of the oil and gasproperties that is audited for both FC and SE firms Therefore, the reliability of the write-down amount should be similar for recognizing and disclosing firms

The SEC required FC firms to recognize a write-down if the net value of

the oil and gas properties exceeded the discounted cash flow from proved oil

and gas reserves The accounting practitioners, before SFAS 121, required SE

8

firms to recognize a write-down if the net value of the oil and gas properties

exceeded the undiscounted cash flow from proved oil and gas reserves The FASB adopted this practice in SFAS 121, stating that “The board adopted the recoverability test that uses the sum of the expected future cash flows

(undiscounted and without interest charges) as an acceptable approach for identifying when an impairment loss must be recognized In many cases, it may be relatively easy to conclude that the amount will equal or exceed the carrying amount of an asset without incurring the cost of projecting cashflows.”

Therefore, choosing the undiscounted value as the threshold was a cost-benefit tradeoff and not a reliability issue The cost-benefit tradeoff does not apply to the oil and gas industry, since the SEC requires the projection of the cashflows for both the SE and FC firms Therefore, in the oil and gas industry investors arenot expected to consider the estimate of a recognized write-down to be more reliable than an estimate of a disclosed write-down

3.2 Time line of recognition and disclosure Except two firms the

recognized write-downs are released at the earnings announcement date The release includes current and last year’s earnings per share and sales In

addition, the earnings announcements will include a separate line specifying the amount of the non-cash write-down included in the current year’s earnings per share The subsequent release of the financial statements provides a

variety of additional information including the cash flow statement, drilling activities, reserves, and the causes of the write-down For SE firms the earnings

9

announcement reports the current and last year’s earnings per share and sales with no mention of the “as-if” write-down The subsequent release of financial statements will include the same information as for the FC firms and the

supplementary unaudited information allows investors to calculate the

economic loss due to the reduction in the oil and gas prices Therefore,

different pricing of recognition and disclosure might be due to a larger

information set at the 10-K filing compared with the earnings announcement The information set at the earnings announcement and the 10-K filings is

discussed in section 3.4

3.3 Economic differences driving the recognition and disclosure choice In cases where recognition results from managerial discretion,

recognition at a certain point in time might be informative for the users of

financial statements This is not so in this study, since managers’ discretion is limited to the FC or SE accounting choice If the managers choose the FC

method they must recognize a write-down if the ceiling is below the net

capitalized costs

Since the accounting choice dictates whether a write-down is recognized, managers of oil and gas firms may choose the method that has the “best” fit fortheir operations, financing, and investment activities For example, Malmquist [1990] documents that SE firms tend to be larger, obtain lower debt levels, and devote fewer resources for exploration and drilling The previously documentedeconomic differences may cause investors to interpret differently the

10

implications of recognized and disclosed write-downs on the firm’s future cash flows Consequently, section 3.4, details the variables that control for the

expected pricing difference between recognizing and disclosing firms

3.4 The explanatory variables To test the pricing implication of recognition and disclosure, I estimate a cross-sectional regression at the

recognized write-downs announcement date and at the 10-K filing date The dependent variable is the 3-day stock return surrounding both events and is

identified using the Nexis Allnews library for the write-down announcement and using the Access Disclosure Database for the 10-K filing date The analysis

includes the day prior to the announcement to control for leakage of news The trading day following the announcement is included to capture investors’

reactions to announcements made after trade ends.6

The regression independent variables are as follows

Unexpected earnings (UE) and unexpected operating cash flows (UOCF)

Since all but two firms announced their write-downs simultaneously with

earnings, a different financial performance across the two groups will cause a different price reaction at the write-downs announcements To control for

financial performance, I calculate the earnings surprise, UE, as the difference between the prior year’s and current year’s net income deflated by the firm’s market value at the current fiscal year’s end For firms recognizing the write-

6 Changing the event window to include only the write-downs announcement day and the following trading day does not change this paper’s results.

11

down, the current year’s net income excludes the write-down.7 Since earnings are known at the filing of the 10-K, the change in operating cash flows at the 10-K filing date is used as the financial performance measure The variable is the current year operating cash flows minus the prior year’s operating cash flows divided by the firm’s market value at the current fiscal year’s end.

Recognized down (WTFC) and as-if down (WTSE) At the

write-down announcement date I use only WTFC WTFC is a slope dummy with a positive value of the pretax recognized write-downs for FC firms and zero for thedisclosing firms At the 10-K filing date, I add WTSE as an explanatory variable WTSE is a slope dummy with the positive value of the pretax disclosed write-downs for SE firms and zero for FC firms Both WTFC and WTSE are deflated by the market value at the fiscal year end However, the recognized write-down is associated with the fourth quarter, while the disclosed write-down may occur anytime during the year Therefore, I adjust the as-if write-down by multiplying

it by the proportion of the fourth-quarter oil and gas price decline out of the annual decline.8

Contractual costs (DCVIOL): the market reaction to the recognized or

disclosed write-down may reflect a change in contractual costs A common measure for contractual costs is the firm’s debt to equity ratio This measure

7 The analyst forecast is not used to form the earnings expectation, since we do not know whether analysts predict the recognized write-down and how accurate they are.

8 The proportion of the decrease in the oil and gas prices attributed to the fourth quarter is (P t-Pt-1) / (Pt-Pt-4), where Pt is the weighted average of oil and gas prices at the end of the current fiscal year, Pt-1 is calculated

at the end of the third quarter, and Pt-4 is calculated at the end of the previous year If all the drop in oil and gas prices occurs in the fourth quarter this calculation is set to one and if none of the drop occurred in the fourth quarter it is set to zero The mean (median) percentage of oil and gas decline that is attributed

to the fourth quarter decline in prices is 0.8441 (1.0)

12

has limitations as firms in the oil and gas industry generally have debt

covenants based on balance sheet financial ratios and off balance sheet value

of oil and gas reserves (for example see Frost and Bernard [1989]) Moreover, prior literature, documents a higher debt to equity ratios for FC firms compared with SE firms That does not necessarily translate to a higher probability of debtcovenant violations for those firms

To control for contractual costs, a variable that captures how the

information reported in the write-downs announcement date effects the

probability of debt covenant violation (DCVIOL) is constructed as follows: for each firm, the off and on balance sheet financial ratios or amounts that are used

in the debt covenant are identified (termed Fm) Subsequently, assuming

normality, the probability of the debt covenant violation for the last period PRt-1

= (Ft-1 - Fm) / σf is calculated, where Ft-1 is the financial variable’s amount or ratiobefore incorporating the information contained in the write-downs

announcement, and σf is the standard deviation of that variable.9 Using the information available at the write-downs announcement date, PRt = (Ft - Fm) / σf

is calculated, resulting in this period’s probability of the debt covenant violation (Ft is the financial variable’s amount or ratio after incorporating the information contained in the write-downs announcement) DCVIOL is the difference

between PRt and PRt-1 In a case where a firm has several debt covenants, the most binding covenant is chosen for the calculation If the write-downs

announcement contains no information on the new value of the financial

9 if the financial variable is available annually the last 10 years are used, if it is available quarterly the last

24 quarters are used.

13

variable, DCVIOL is set equal to zero DCVIOL contains considerable estimation error at the write-downs announcement date compared with the 10-K filings, because of the small amount of information provided in the write-downs

announcement.10 This calculation is repeated for the probability of debt

covenant violation at the 10-K date (DCVIOL10K) At the 10-K date, I can use the footnote information that contains the value of the financial variables and a discussion of whether the debt covenants changed Therefore, DCVIOL10K includes the change in the probability of debt covenant violation for the period between the write-downs announcement and the 10-K filing date

Industry concentration (INDCON and INDCON10K) The market reaction to

recognition and disclosure may be affected by the firms’ differential

involvements in oil and gas production For example, Zacks Investment

Research recommends excluding from EPS a write-down of oil and gas

properties by a firm that is not primarily engaged in the oil and gas business Tocontrol for the firms’ concentration in the oil and gas industry, I use the

percentage of total revenue due to oil and gas operations At the write-down announcement, since the revenue attributed to oil and gas sales is unknown, INDCON is the previous year’s oil and gas sales divided by total sales, and at the 10-K filing INDCON10K is the current year’s oil and gas sales divided by the current year’s total sales

10 For example, a debt covenant specifies that a firm must keep reserves at a level of $25 million At the write-downs announcement, the size of the actual write-down is known Assuming that the net oil and gas properties value and the debt covenants did not change, I can conclude that the reserves decreased by at least the write-down size I have no new information for firms only disclosing the loss at the earnings announcement Only at the 10-K date will I have sufficient information to calculate the exact effect of the write-down on the debt covenant violation probability.

14

Size (MV) In the last 20 years, several anomalies related to firm size

have been documented in various accounting and finance studies For

example, Ball and Kothari [1991] document that surrounding earnings

announcements, abnormal returns are positive and decreasing in firm size Therefore, MV is the controlling variable for firm size and is calculated as the natural log of the firm’s market value at its fiscal year end for the year the recognized or disclosed write-downs are identified

The 10-K filing includes a large volume of information The information for

SE firms might be favorable and could “cancel” the negative implications of the economic loss Although I attempt to control for the information contained in the 10-K, no two firms are identical.11 Moreover, since the market reaction

depends on investors’ expectations, identical information can result in different market reactions Therefore, two competing hypotheses can explain a different market reaction to recognition versus disclosure One hypothesis is that

investors interpret the valuation implications of information contained in

footnotes differently than information formally recognized in the primary

financial statements Alternatively, the 10-K contains favorable information thatnullifies the bad news contained in the disclosed loss.12

11 In the robustness tests section several explanatory variables are included in the 10-K filing Among others they include changes in reserve value and quantity, drilling and exploration activity, and sale purchase and acquisition of reserves in place.

12 If we assume that on average the financial statements for SE and FC firms contain the same proportion of favorable information, there is no problem in conducting the test However, this is quite a strong assumption to make.

15

4 Sample Selection and Descriptive Statistics

4.1 Sample selection The sample of recognized write-downs is based on

financial statement footnotes for 1987-1994 from both LEXIS and Corporate Text The period was chosen because both sources of information contain 10-Ks

beginning in 1987 The study includes only firms that recognize or disclose write-downs due to applying the SEC ceiling test

Table 1, panel A contains the quarterly distribution of recognized downs for the 1987-94 period Using the two above sources of information, 39 firms with 60 recognized write-downs are found.13 Since footnote disclosure, required for estimating the as-if write-down is available only in annual reports, this study includes only the 26 recognized write-downs that occurred in the

write-fourth quarter of fiscal years 1990-93 Next, for each firm the Nexis library is

searched for the first mention of the write-down.14

To identify firms disclosing write-downs, all the firms that Compustat classifies as SE method firms (Footnote 31 in Compustat) between 1990 and

1993 are identified The period 1990 to 1993 was chosen because of the effort required to analyze the footnotes Moreover, between 1988 and 1990, oil pricesincreased (gas prices remained steady), thus reducing the chance that SE firms will disclose write-downs Subsequently, the available financial statement footnotes from 1990 to 1993 for the sample of SE firms are analyzed

13 The search keywords are: “red! of carr!”, “imp! Of carr!”, ”wri! Of carr!” And “oil and gas” Homestake Mining which recognized a write-down in the fourth quarter of 1993, was deleted from the sample Homestake Mining was dropped since it is a gold producing firm with one oil well, and therefore is not required to provide any footnote information on the ceiling calculation.

14 For non fourth quarter write-downs 2 firms released a warning of the write-down before the earnings announcement.

16

Compustat reports about 230 SE firms between 1990 and 1993, while the

number of financial statements used in this study varies from 101 in 1990 to 95

in 1993 The reduction in sample size is mainly due to the elimination of

utilities and missing financial statements.15

Table 1, panel B provides the number of SE firms that disclosed a down The total number includes 46 disclosed write-downs and four recognized write-downs The 50 write-downs are composed of 30 firms having one requiredwrite-down and 10 firms with two required write-downs.16 The proportion of “as-if” write-downs for SE firms is 12.43 percent (fifty out of 402 financial

write-statements) compared with 9.54% for firms recognizing a write-down (forty-fourout of 461).17

4.2 Descriptive statistics for the sample Descriptive statistics for the

sample for 1989-93 are presented in table 2 Results of applying a t-test,

median test, and Wilcoxon non-parametric test, indicate that the financial

performance of FC firms is insignificantly different from that of SE firms When the recognized write-down is ignored, both groups display the same change in earnings In addition, the change in operating cash flows and the fiscal year

15 Firms listed as ADR and firms concentrating in gold mining were also eliminated Finally, Compustat reports 103 SE firms with total assets of less than $2 million Financial statements for only one such firm (Oil City Petroleum) were found However, no price data for this firm was available on CRSP.

16 Since the analysis is perform year by year, a total of 79 write-downs exist However, twenty-nine write-downs are eliminated due to double counting that can be illustrated by the following example Assume that in period one firm X has $100 in net capitalized costs and the ceiling is $90, therefore requiring an estimated write-down of $10 Assume that in period two the net capitalized costs are $85 and the ceiling is $80 On a year by year basis a write-down of $5 is required in period two However, the firm’s net capitalized asset

is adjusted to $75 in period two ($85 minus the $10 “as-if” write-down in period one) such that no write-down is required in period two.

17 The number of FC firms in Compustat from 1990-93 is 841 However, for comparison reasons 89 firms with total assets less than $2 million, and 309 utilities, gold mining, and foreign firms are eliminated.

17

end market value are insignificantly different for the two groups.18 Finally,

although the FC firms debt to market value ratio is more than double that of the

SE firms, both groups display an increase in the probability of debt covenant violation (Δdcv).dcv).19 Moreover, out of 24 SE firms with debt covenants 16 had an increase in Δdcv).dcv compared with 15 out of 17 for FC firms Since debt covenantsare written on balance sheet ratio and off balance sheet reserve information, it

is reasonable that recognizing firms have a higher proportion of firms with an increase in the probability of debt covenant violation

In addition, table 2 shows that FC and SE firms significantly differ in their operations FC firms are significantly less diversified (measured as the

percentage of oil and gas sales out of total sales) and devote more resources toexploration and drilling.20 In addition, the success rate of drilling exploratory and development wells (not reported in the table) is significantly higher for FC firms compared with SE firms Moreover, SE firms that have larger drilling

portfolios do not drive the result These results substantiate the observation that FC firms concentrate more on drilling and exploration and less on

production

Table 3, panel A provides the market reaction for the three days

surrounding the write-down announcement for recognizing firms.21 The table

18 The change in sales is also insignificantly different between the two groups.

19 The debt to market value ratio is not reported in the paper For FC firms, the mean is 97% and median is 88%, compared to a mean of 47% and median of 33% for SE firms The mean (median) probability of debt covenant violation (before the inclusion of the current year information) for FC firms is 0.2265 (0.063) and 0.178 (0.0) for SE firms The mean and median are insignificantly different from each other.

20 When seven integrated SE firms are removed, both groups display the same diversification.

21 All the results in this paper are based on raw returns Repeating the analysis with CAPM adjusted returns and size adjusted returns yielded consistent results

18

includes 51 FC firms and nine SE firms recognizing a write-down between 1987 and 1994 Overall a negative market reaction to the write-downs is

documented This is consistent with studies conducted by Elliott and Shaw (1988) and Strong and Meyer (1987) that document a significant negative

market-reaction to write-down announcements Moreover, a negative market reaction is not surprising since the write-down relates to the firms’ core

business, affecting current sale revenue per unit However, only the FC firms exhibit a significant negative market reaction FC firms write down their assets

to the discounted value while SE firms write down their assets to the

undiscounted value Consequently, the mean (median) write-down deflated by the market value of equity for FC firms is 0.376 (0.217) and for SE firms 0.088 (0.080) In addition, the correlation between the write-down size and the stock price return is -0.407 (p-value of 0.0009) One plausible explanation is that investors classify, at the write-downs announcement date, a small write-down

as transitory and a large write-down as permanent.22

In addition, table 3 panel B reports the market reaction by quarter Most

of the write-downs occur in the fourth quarter Interestingly, only the fourth quarter write-downs returns are significantly negative However, except for the second quarter, the correlation between the write-down size and the stock pricereturn is significant at the 15% level (at the second quarter the p-value is

0.914) Therefore, this result may be due to the small sample of write-downs in the first three quarters compared with the fourth quarter

22 The monthly standard deviation of oil price changes between 1989 and 1994 was 0.104 (0.1263 for the gas prices).

19

5 Results and Diagnostic Tests

5.1 Regression results To investigate the effects of recognition and

disclosure, The following cross-sectional regression is estimated at the downs announcement date:

write-RETit = α + β1UEit +β2WTFCit + β3DCVIOLit + β4INDCONit +β5MVit + εit (1)

where the variables are defined in Section 3.4

The results of estimating the regression are presented in table 4, panel A The test excludes one FC firm that is an outlier and two SE firms with no

information on their previous year’s oil and gas operations.23 The coefficient on the recognized write-down amount (WTFC) is -0.061 with a t-statistic of -3.75 This is consistent with the univariate results which document a negative return for FC firms recognizing a write-down at the fourth quarter In addition, the coefficient on DCVIOL the variable that captures the increase in the probability

of debt covenant violations, is negative as expected with a t-statistic of -2.2 Moreover, the coefficient is negative for both the SE and FC firms For FC firms the coefficient is -0.0596 (t-statistic of -1.91) and for SE firms it is -0.061 (t-statistic of -1.12) The coefficient on UE is 0.008 with a t-statistic of 2.34

However, the coefficient on the UE of firms recognizing the writedown is

-0.0006 (t-statistic of -0.03), while the coefficient on UE for firms disclosing the

23 The FC firm, Falcon Oil and Gas displayed a 66% stock price decline at the announcement of the down and had a r-student value of over 6 Including the firm in the regression considerably increased the t-statistic on WTFC

write-20