*460 Amount of tip income received in 1978 by cocktail waitresses serving complimentary drinks on the casino floor of Resorts International Hotel and Casino in Atlantic City, New Jersey, determined.
SWIFT, Judge: In these consolidated cases, respondent determined the following deficiencies in and additions to petitioners' Federal income tax liabilities for the year 1978 as follows:
| Additions to Tax | |||
| Docket No. | Tax | § 6653(a) 2 | |
| Theresa J. Bruno | 9578-82 | $2,292.00 | $114.60 |
| Cathy A. Stutzman | 11000-82 | 3,271.00 | 163.55 |
| Margo A. Achille | 13439-82 | 4,655.74 | 232.79 |
| Bonnie L. Gasperini | 13484-82 | 1,031.00 | 51.55 |
| Clara M. Fesco | 27315-82 | 3,642.00 | 182.00 |
The issues for decision herein are: (1) Whether petitioners understated tip income on their respective 1978 Federal income tax returns, and if so, by what amounts, and (2) whether petitioners are liable for the 5-percent addition to tax under section 6653(a). In the main body of our findings of fact and opinion we address matters pertinent to all petitioners. In an appendix, we have made*462 additional findings of fact pertinent to each separate petitioner.
FINDINGS OF FACT
Some of the facts have been stipulated and are so found. At the time the petitions herein were filed, each of the petitioners resided in the state of New Jersey.
During the last six and-a-half months of 1978, petitioners were employed as cocktail waitresses at the Resorts International Hotel and Casino in Atlantic City, New Jersey (hereinafter referred to as "Resorts"). In that capacity, they spent a major portion of their time serving complimentary drinks to gamblers on the casino floor. The Resorts hotel and gambling casino opened to the public on May 24, 1978. Resorts was the first licensed gambling casino to open in Atlantic City, after enactment in New Jersey of laws legalizing casino gambling. Resorts also was the only casino in operation in Atlantic City in 1978.
Each petitioner timely filed an individual Federal income tax return for 1978. The amount of tip income separately reported on petitioners' respective 1978 returns ranged from zero by petitioner Achille to $3,200 by petitioner Gasperini. None of the petitioners maintained or kept contemporaneous records of tip income received*463 in 1978.
The work performed by petitioners and Resorts' records
As cocktail waitresses assigned to Resorts' casino floor, petitioners' responsibilities included taking orders for, obtaining, and serving, complimentary drinks (both alcoholic and non-alcoholic) to gamblers on the casino floor. Each waitress was assigned a particular area or "pit" on the casino floor. Each pit consisted of a designated group of tables in a particular gaming area (either "twenty-one", roulette, craps, baccarat or "big wheel" 3). The waitresses served complimentary drinks only to gamblers actually placing bets at the tables. Individuals merely observing the action were not to be served complimentary drinks.
After taking orders for drinks, each waitress would walk to the service bar, obtain glasses, place ice in the glasses, and then would wait for the service bartender to mix or pour the drinks ordered. The waitress next would walk back to her pit, find the gamblers who had ordered the drinks and serve them the drinks. A barboy assigned to each shift assisted the waitresses by putting out glasses, cutting garnishes, *464 and making sure the ice machine was working.
Petitioners normally worked six-hour shifts, five days per week. There were three shifts per day as follows: Day shift, 10 a.m.-4 p.m.; swing shift, 4 p.m.-10 p.m.; and graveyard shift, 10 p.m.-4 a.m. On Fridays, Saturdays, and holidays the graveyard shift would last until 6 a.m. Petitioners were allowed one scheduled half-hour break per six-hour shift, during which time a substitute or "breaker" would cover their assigned pits, and additional breaks of five minutes per hour as needed. In return for serving drinks, petitioners were paid wages by Resorts at a level below the minimum wage. 4 They did, however, also receive tips from gamblers for serving the complimentary drinks.
Because Resorts was the first and only gambling casino open on the East Coast in 1978, *465 the casino floor during that year was frequently very crowded. It was not unusual to have three or more individuals waiting behind each individual actually placing bets for their opportunity to play. In addition, in 1978 there were a significant number of people who came to the casino floor out of curiosity, merely to watch others play, because they had not previously been in a gambling casino.
Occasionally in 1978, Resorts had to initiate crowd control measures in order to limit the number of people on the casino floor. Although the large number of people on the casino floor in 1978 constituted a large source of potential tips for the cocktail waitresses, the crowded conditions also made it difficult for the waitresses, with drinks and trays in hand, to walk quickly to and from their pit areas to obtain and serve drinks. Occasionally, it was even necessary for the waitresses to join arms with a security guard in order to successfully maneuver their way through the crowds.
Resorts encountered a number of start-up problems during 1978. Significant to this case is the fact that the procedures used by Resorts to maintain an accurate inventory of the alcoholic beverages in that*466 year were not accurate. The inventory records of the liquor purchased were not carefully maintained.It was not unusual to find several bottles of liquor missing from the service bar area at the beginning of a shift. Except for December, the bartenders at Resorts utilized the "free-pour" method of dispensing alcoholic drinks into the glasses throughout 1978, with the frequent result that a larger portion of liquor was poured into the average drink than the 1.25 ounces estimated by Resorts for purposes of its inventory and bookkeeping procedures. Double and triple shots of liquor often were poured into one glass by the bartenders.An automatic measuring "gun" system of dispensing liquor, which would dispense only the specified amount of liquor per drink, was not instituted until December of 1978.
An additional start-up problem material to this case was that the 1978 payroll records were not carefully maintained. These records for 1978 reflected significant errors and inconsistencies in the number of hours worked by the cocktail waitresses.
The IRS Audit
In 1980 and 1981, food and beverage servers employed by the hotels and gambling casinos located in Atlantic City, New*467 Jersey, were the subject of an Internal Revenue Service coordinated audit project referred to within the Internal Revenue Service as "The Atlantic City Tip Project." One important part of that project was to establish average hourly rates of tips that could be used in auditing the 1978 Federal income tax returns of the food and beverage servers employed by the hotels and casinos in Atlantic City in 1978.
In the case of food and beverage servers who worked in the restaurants and bars within each hotel and casino (as distinguished from those who worked on the gambling casino floors), the methodology used by the Internal Revenue Service audit team to determine average hourly rates of tips was as follows. Credit card sales slips for a sample test period for each restaurant and bar were used to determine the average tips received as a percentage of total food and beverage credit sales (generally taking into account a "stiff" rate ranging from zero to 40-percent). That percentage was applied to the total credit and cash sales of each restaurant and each bar in order to determine an estimated total amount of gross tips received by all of the food and beverage servers at a particular restaurant*468 or bar during 1978. The total estimated amount of gross tips for each restaurant and bar then was divided by the total number of hours collectively worked by all of the food and beverage servers in each restaurant and bar in order to determine an average tip rate per hour for food and beverage servers at each restaurant and bar location.
Depending on the practices of each particular restaurant and bar (as determined in interviews), a further adjustment where appropriate would be made in the average tip rate per hour for the sharing of tips with busboys and bartenders. In reconstructing the tip income of a particular food or beverage server for any one year, the average net tip rate per hour that was calculated in the manner described above for each restaurant and bar would be multiplied by the number of hours worked by each food and beverage server in each restaurant and bar location.
Using the above methodology, respondent calculated that the average net tip rates per hour for food and beverage servers at the various restaurants and bars located within Resorts during 1978 (not including the cocktail waitresses who served complimentary drinks on the casino floor) were as follows: *469
| Restaurant and Bar Locations | Average Net | Percentage of |
| within Resorts (excluding Casino | Tip Rate | Net Tips to |
| Floor) | Per Hour | Gross Sales |
| Le Palais (gourmet restaurant) | $5.80 | 14.64 % |
| Wedgewood Pavillion (informal | ||
| restaurant) | 5.53 | 11.84 % |
| Camelot Restaurant | 6.23 | 14.03 % |
| Room Service | 2.43 | 15.86 % |
| Superstar Theatre (showroom) | 4.20 | 16.21 % |
| Renaissance/Gold Room (restaurant) | 3.92 | 10.75 % |
| Rendezvous Lounge (bar) | 7.12 | 13.42 % |
| Disco Bar | 8.35 | 17.16 % |
| Gallery Bar | 5.56 | 15.23 % |
Petitioners herein, some of whom spent a portion of their time working in several of the locations listed above, in addition to serving complimentary drinks on the casino floor of Resorts, do not challenge any of the average net tip rates per hour as calculated by respondent for the time they worked at the above locations.
What is contested by petitioners herein is the average net tip rate per hour calculated by respondent for cocktail waitresses who served complimentary drinks on the casino floor. That average was calculated by respondent to be $24.91 per hour, or (as a comparison with the average net tip rates in the above list will indicate) from three to*470 ten times the average tips per hour as determined for food and beverage servers at all other locations within Resorts.
The methodology used by respondent to calculate his figure for the average net tip rate per hour to be applied to the waitresses serving complimentary drinks on the casino floor was significantly more complicated than the methodology used for the calculation of average tip rates for the other locations. The complication arose primarily from the absence of actual liquor sales figures on the casino floor, which was due to the complimentary nature of the drinks, and due to Resorts' failure to maintain accurate records of the volume or number of complimentary drinks served.
The steps used by respondent to determine his estimated tips per hour for the cocktail waitresses serving complimentary drinks on the casino floor are summarized in the schedule set forth below. For each of the seven steps of respondent's calculation, we have also described in the summary schedule the source from which the information used by respondent was obtained, and we have set forth the results reached by respondent at each step.
| RESULTS | ||
| STEPS-- | OBTAINED BY | |
| CALCULATION OF | SOURCE OF DATA | RESPONDENT |
| (1) AVERAGE TIP | ||
| PER DRINK | I.R.S. Surveillance | $ .81 per drink |
| (2) TOTAL DRINKS SERVED | Imputed sales-- | |
| ON CASINO FLOOR | Resorts' Records | 1,590,807 |
| (3) TOTAL TIPS | Arithmetic--Tips per | |
| Drink times Total Drinks | ||
| ($ .81 times $1,590,807) | $1,288,544 | |
| (4) TOTAL HOURS | Resorts' payroll records | 51,719 |
| (5) TIPS PER HOUR | Arithmetic--Total Tips | |
| divided by Total Hours | ||
| ($1,288,554 divided by | ||
| 51,719) | $24.91 | |
| (6) HOURS WORKED BY | Different | |
| EACH COCKTAIL | for each | |
| WAITRESS | Resorts' payroll records | waitress |
| (7) TIPS RECEIVED BY | Arithmetic--Tips per Hour | Different |
| EACH COCKTAIL | times Hours Worked by | for each |
| WAITRESSES | each cocktail waitress | waitress |
*471 We now will discuss separately and in more detail each step of respondent's calculation, and the manner by which respondent developed the information used at each step.
Step 1: Average Tip Per Drink: The I.R.S. Surveillance
As is indicated in the above schedule, respondent used surveillance data only to calculate a figure for the average tip per drink. Respondent did not use data obtained from the surveillance to calculate a figure for the average number of drinks per hour, even though that figure was available from the surveillance data obtained.
The average tip per drink for the cocktail waitresses serving complimentary drinks on the casino floor was determined through an extensive surveillance program conducted at Resorts and other Atlantic City casinos in 1980 and 1981 by undercover special agents of the Criminal Investigation Division of the Internal Revenue Service. The purpose of the surveillance was to determine a figure for the average tip per drink by observing the number of drinks served by the cocktail waitresses and the amount of tips received, if any, during designated observation periods at designated gambling pits within a given casino.
Each of the*472 special agents had received at least two days of surveillance training, and each also had participated in some prior surveillance work. At orientation meetings prior to each of the surveillance periods in 1980 and 1981, instructions were given to the agents as to where, when and how to conduct the surveillance. Copies of the floor plans of the casinos that were to be subject to the surveillance and written descriptions of the colored chips used at the various casinos were given to each agent.
The 1980 surveillance program was conducted during only a two-week period in December, 1980, and involved only two casinos--Resorts and Ceasar's Palace Hotel and Casino. The undercover agents worked in groups of three and were instructed to observe the number of drinks served and the tips actually received by each cocktail waitress for 15-minute periods out of designated half-hour periods. Each 15-minute period was to begin when a cocktail waitress first entered a specified pit area after the start of each designated half-hour period. The 1981 surveillance program was much more extensive, extending over a six-month period from July to December of 1981 and covering seven casinos. For the 1981*473 surveillance, agents worked in groups of two and were instructed to observe a specified pit area for the entire 30 minutes of each designated half-hour period. With the exception of these differences, the procedures and instructions were identical for the surveillance programs in 1980 and 1981.
Prior to beginning a surveillance on any particular day or night at each casino, the agents were given time and pit assignments within the casinos based upon random selections generated by a computer. Each agent would then meet with his or her other team member or members and proceed to the designated pit. Several alternate pit assignments were provided in case the designated pit was not open at the specified time to begin the surveillance. The agents would orally confer and would divide the tables in the designated pit area between themselves. During each fifteen minute or half-hour observation period, each agent would mentally note the number of drinks served to gamblers and the tips received by each cocktail waitress in the particular pit under surveillance. The agents were instructed that if they saw something being given to a waitress but could not tell what it was, they were to treat*474 it as if no tip had been received. If they could tell that a chip was given to a waitress but could not tell what color the chip was, they were to treat it as a tip of $1.00 (the minimum chip value for a valued chip). If they could tell that a coin was given to a waitress but they could not clearly determine the value of the coin, they were to treat it as a quarter, and if currency was given to a waitress but the amount thereof could not be seen, they were to treat it as a $1.00 tip.
Immediately following each fifteen minute or half-hour observation period, the two or three agents in each team left the pit area under surveillance and met to discuss and resolve any problems that they may have experienced in their surveillance during that particular surveillance period. Each agent then filled out a 3 by 5 index card and recorded the number of drinks and the amount of tips observed for each surveillance period. Each agent would initial each index card of the other agents on his team. The cards were then collected by the surveillance coordinator and summaries were prepared of the information on the cards.
Using only the information obtained from the 1980 and 1981 surveillances,*475 respondent's statistical expert determined that the average tip per drink was approximately $ .81 in 1980 and $ .83 in 1981. In his notices of deficiency to petitioners, respondent utilized the lower figure of $ .81 as the average tip per drink for purposes of his subsequent computations since that was more conservative than the $ .83 per drink tip figure derived from the 1981 surveillance.
At trial, petitioners' and respondent's experts both accepted the basic statistical validity of the 1981 surveillance in terms of the sampling size and standard deviations. However, they disagree slightly as to the amount of a reduction in the average tip rate that is appropriate to increase the confidence they have in the results of the survey. Petitioners' expert would use a "lower confidence limit test" 5 of 97.5-percent (reducing the tip rate per drink to $ .75), and respondent's expert would use a confidence limit test of 95-percent (reducing the tip rate per drink to $ .77).
*476 Petitioners' statistical expert also disagrees strongly with the use by respondent of any tip per drink rate derived from the surveillance in combination with the other numbers derived by respondent from Resorts' books and records, as explained further below, and he also questions whether data obtained in 1980 and 1981 should properly be used to draw conclusions regarding the year 1978.
Step 2: Total Drinks: Imputed Sales Figures
After establishing through the surveillance program conducted in 1980 and 1981 a figure of $ .81 as the figure he would use for the average tip per drink, the next step in respondent's calculations was to determine an estimate of the number of drinks served during the last six months of 1978 by the cocktail waitresses on Resorts' casino floor. As mentioned earlier, respondent did not use the surveillance data to make this determination, but rather utilized what information pertaining thereto he could obtain from Resorts' records and files for the last six months of 1978.
First, using the rough internal cost accounting or bookkeeping charges of Resorts for 1978, a cost of $373,692 was determined for the liquor charged to the two service bars*477 on the casino floor for the months of July through December of 1978.That cost figure was grossed up by respondent to obtain a cost figure for all beverages, both alcoholic and non-alcoholic. For the gross-up calculation, an assumed cost ratio of alcoholic to non-alcoholic drinks was used of .936 to .064, and a total cost figure for both alcoholic and non-alcoholic drinks served on the casino floor was determined of $399,244 for the months of June through December of 1978.
Respondent then divided the $399,244 total estimated cost of alcoholic and non-alcoholic drinks by a cost factor of .14 in order to calculate an imputed sales figure for all drinks of $2,851,743, a calculation that assumes a 714-percent mark-up in the cost of drinks to the imputed sales figure. Respondent apparently obtained the cost factor of .14 from an employee who worked in the inventory and cost accounting department of Resorts. Where the employee obtained that figure and the basis therefor were not explained.
From the total imputed sales figure calculated by respondent of $2,851,743, for both alcoholic and non-alcoholic complimentary drinks served by Resorts on its casino floor in the last six months of*478 1978, respondent subtracted $2,328,612 as an estimate of the imputed sales of alcoholic drinks and thereby arrived at an estimate of $523,131 for the estimated imputed sales of non-alcoholic drinks.
Each non-alcoholic drink was assigned an estimated sales value by Resorts of $1, and respondent therefore divided the total $523,131 estimated imputed sales value of the non-alcoholic drinks by 1 to determine his figure for the estimated number of non-alcoholic drinks served during the last six months of 1978 (namely, 523,131). A further adjustment to the estimated 523,131 number of non-alcoholic drinks was made by respondent to take into account the volume of non-alcoholic beverage that is not served as a separate drink but that is served as a mix with liquor. Respondent estimated that 24-percent or 125,551 of the total estimated non-alcoholic drinks were actually mixed with liquor. Therefore the net estimated number of non-alcoholic drinks served by Resorts on the casino floor was calculated by respondent to be 397, 580 for the period of June through December of 1978 (523,131 minus 125,551 equals 397,580).
After making the above calculations in order to calculate his estimate of*479 the number of non-alcoholic drinks served in 1978, respondent calculated an estimate of the number of alcoholic drinks. To make that calculation, respondent did not use the $2,328,612 estimate of imputed sales of alcoholic drinks that he used in the prior calculation. Had he used that figure and divided it by $2 per drink (the imputed sales value for each alcoholic drink), respondent would have obtained a figure of 1,164,306 for the number of estimated alcoholic drinks served during the last six months of 1978.
Instead, respondent used a higher figure of 1,409,030 for his estimate of the number of alcoholic drinks served during that period, a figure obtained from Resorts' 1978 Complimentary Services Schedule and purportedly representing the number of complimentary alcoholic drinks served during 1978. Respondent then made an adjustment for drinks actually served by the waitresses in locations other than the casino floor but charged to the service bars on the casino floor of 11,254, and respondent made another adjustment for 204,549 drinks supposedly served prior to July 1, 1978, (based on an assumption that because 14.634-percent of the costs incurred by the casino service bars occurred*480 prior to July 1, 1978, the same percentage of alcoholic drinks, or 204,549 drinks, were served prior to that date).
With the reductions mentioned in the prior paragraphs, respondent determined that an estimated 1,193,327 alcoholic drinks and 397,580 non-alcoholic drinks were served on the casino floor of Resorts during the last six months of 1978, for a total of 1,590,807 drinks.
Step 3: Total Tips: Arithmetic
After calculating an estimate of the average tip per drink and of the total drinks served during the last six months of 1978, respondent simply made the arithmetic computation of multiplying his determination of the average tip per drink ($ .81) by his figure for the estimated total drinks (1,590,807) to obtain an estimate of the total tips received during the last six months of 1978 by all of the cocktail waitresses who served complimentary drinks on Resorts' casino floor. That estimated total tip figure calculated by respondent was $1,288,554.
Step 4: Total Hours: Resort's Payroll Records
Having completed his calculation of the estimated total tips received during the last six months of 1978, respondent next sought to determine the total hours worked*481 by the cocktail waitresses on Resorts' casino floor during that same period of time. Respondent's agents made a study of Resorts' payroll records for the last six months of 1978 and determined that the cocktail waitresses assigned to the casino floor worked a total of 51,719 hours during that period. To arrive at that figure, respondent's agents analyzed Resorts "gross pay proofs" and used Resorts' job description codes to identify individual cocktail waitresses who worked on the casino floor and the number of hours worked at that location. All hours worked by waitresses on the casino floor were than added up.
At trial, several glaring errors in Resorts' payroll records for 1978 were established. For example, in 1978 petitioner Bonnie Gasperini received a pay stub (which should contain the same information as the pay proofs for Ms. Gasperini for the same pay period), that indicates she worked 113 hours in one particular week. That same pay stub also reflected that she received pay for that week of only $22.60. Because of the fact that the agents who conducted this aspect of the tip income project did not keep any work papers, adding machines tapes or other records, it is impossible*482 to determine whether or not obvious mistakes such as the error on Ms. Garperini's pay stub were eliminated in the figures developed by respondent's agents.
There is also evidence of less obvious errors in Resorts' payroll records. For example, those records indicate that Cathy Stutzman, a petitioner herein, worked 134 hours in the Rendezvous Lounge and 158.5 hours in the Gold Room during 1978. However, Ms. Stutzman testified at trial (even though such testimony was against her interest given the much higher average tip rate calculated for work on the casino floor versus the other locations) that all of her working hours were spent on the casino floor, and that she never worked in the Rendezvous Lounge or the Gold Room.
Step 5: Tips Per Hour: Arithmetic
Step five of respondent's calculation simply involves the division of the estimated total tips received in 1978 of $1,288,554 by the estimated 51,719 total hours worked by the cocktail waitresses on the casino floor during the last six months of 1978. That division produces a figure for the estimated tips received per hour of $24.91.
Steps 6 and 7: Hours Worked and Tips Received by Each Cocktail Waitress
*483 At steps six and seven, respondent attempts to apply his estimated figure for the average tip rate per hour, to each individual cocktail waitress. He first uses the gross payroll proofs of Resorts to determine the hours worked by each cocktail waitress on the casino floor, and, in step seven, he multiplies that figure for each waitress by his average tip rate per hour of $24.91, to calculate his estimate of the tips received by each waitress for the hours worked on Resorts' casino floor.
As we previously have explained, there were a number of significant errors in Resorts' payroll records for 1978, and where more reliable information was submitted into evidence, we have found it necessary to make certain adjustments to the figure utilized by respondent for the hours and locations worked by some of the petitioners herein in 1978. Those adjustments as to each petitioner are reflected in the appendix hereto. Where no other reliable information is available, we find that respondent appropriately utilized the information contained in Resorts' payroll records as the basis for determining how many hours each cocktail waitress worked in 1978 on the casino floor.
OPINION
It is*484 well settled that tips received are includible in gross income as compensation for services rendered. Section 61(a); section 1.61-2(a)(1), Income Tax Regs.; Killoran v. Commissioner,709 F.2d 31 (9th Cir. 1983), aff'g a Memorandum Opinion of this Court; Roberts v. Commissioner,176 F.2d 221 (9th Cir. 1949), aff'g 10 T.C. 581 (1948). See also Olk v. United States,536 F.2d 876 (9th Cir. 1976), rev'g. 388 F. Supp. 1108 (D. Nev. 1975). Petitioners herein do not contend otherwise. They do contend, however, that respondent's hourly rate of $24.91 used in reconstructing their tip income for the hours they worked as cocktail waitresses on the casino floor of Resorts in 1978 is excessive and that respondent's methodology in calculating that figure is erroneous.
Petitioners contend, based on expert witness testimony they offered, that the correct tip rate per hour in 1978 was $8.40. For the reasons set forth below we find that the appropriate tip rate to be used in reconstructing the 1978*485 income of the cocktail waitresses who served complimentary drinks on Resorts' casino floor is $12.46 per hour.
All taxpayers are required by law to keep records which are sufficient to enable respondent to determine their correct Federal income tax liability. Sec. 6001. Where, as in the situations herein, a taxpayer receives income from tips on a daily basis, this necessarily requires the maintenance of accurate and contemporaneously prepared diaries or records of tip income received, Sec. 1.6001-1(a), Income Tax Regs., and such records must be retained by the taxpayer "so long as the contents thereof may become material in the administration of any internal revenue law." Section 1.6001-1(e), Income Tax Regs; Meneguzzo v. Commissioner,43 T.C. 824, 831 (1965).
If a taxpayer either keeps no records of tip income received, or it appears that the records do not clearly reflect tip income received, respondent is authorized by*486 section 446 to compute income under a method which, in his opinion, does clearly reflect income. Meneguzzo v. Commissioner,supra at 831. In the present case, petitioners either kept no records of their tip income for 1978 or presented "diaries" which were prepared at least three years later. Therefore, respondent attempted to reconstruct petitioners' tip income.
Where a reconstruction of a taxpayer's income is necessary, respondent has great latitude in the method used to make the reconstruction. The method adopted by respondent, however, must produce a result that is reasonable and substantially correct. Mendelson v. Commissioner,305 F.2d 519, 523 (7th Cir. 1962). Although respondent's reconstructions of income need not be exact, they must "be reasonable in light of all the attendant facts and circumstances." Schroeder v. Commissioner,40 T.C. 30, 33 (1963).
The use of an undercover surveillance program in the reconstruction of tip income by respondent has previously been upheld by this Court. In Catalano v. Commissioner,81 T.C. 8 (1983),*487 on appeal (9th Cir., Nov. 25, 1983), an extensive surveillance program was conducted by respondent's agents at Caesar's Palace Hotel and Casino in Las Vegas, Nevada, over a 48-day period from February, 1976, through January, 1977. The subjects of the surveillance were blackjack, roulette and Big Wheel dealers employed by Ceasar's Palace. Respondent's agents covertly observed the exchanges of chips received as tips by the dealers for cash, which exchanges took place at the cashier's cage, located in a public area of the casino. Based upon the surveillance data gathered in that case, respondent calculated an average tip rate of $7.93 per hour for use in reconstructing the dealers' tip income. Respondent's methodology and average tip rate per hour were upheld by this Court.
Under the circumstances of this case, respondent's reconstruction of petitioners' tip income requires some major adjustments. 6
*488 Problems With Respondent's Calculations
The major problem with respondent's calculations is the indirect, artificial, and internally inconsistent manner in which the number of drinks served on the casino floor is developed. In making that computation, the number of non-alcoholic drinks is calculated using some cost data derived from Resorts' records, but also using some highly questionable data apparently obtained informally from Resorts' employees without any explanation or attempt by respondent to corroborate the validity of the data. Such data includes the figure used by respondent for the ratio of alcoholic beverage cost to non-alcoholic beverage cost (.936 to .064), the cost markup figure used to calculate an imputed total sales figure (which resulted in a markup in the imputed sales of alcoholic and non-alcoholic drinks of over 714-percent), the percentage of non-alcoholic beverages used as mixes in alcoholic drinks (24%), and the imputed sales values of complimentary drinks ($2.00 for alcoholic; $1.00 for non-alcoholic). While this data may have been useful for certain cost accounting and other management information purposes, their level of reliability in accurately*489 calculating the actual number of drinks served on Resorts' casino floor, a task for which they were not intended, is insufficient.
A number of witnesses from Resorts testified at trial as to the inaccuracies in Resorts' records for 1978, due to the many start-up problems Resorts was experiencing in that year. The degree to which respondent relied on such suspect records and data informally obtained from Resorts' employees in this case introduces an unacceptable level of uncertainty into respondent's calculations. For example, the figure used by respondent to calculate the number of alcoholic drinks served by Resorts in 1978 was obtained from a form Resorts filed with the State Casino Control Commission for the year 1978. An internal memorandum from the Director of Accounting and Revenue at Resorts (whose duties included food and beverage costs control) stated, with respect to the 1978 form that, "all the figures are very arbitrary and could be developed in other ways if we knew the importance of great accuracy." A memorandum from an assistant manager of the Cost Control Department of Resorts explained that "the actual number of drinks served in the casino may only be 1/2 the number*490 charged to the casino" for internal accounting purposes.
An assumption made by respondent in the part of his calculations which use the "cost mark up" method is that each alcoholic drink contained 1.25 oz. of liquor, the amount of alcohol officially prescribed for each drink by Resorts. However, the testimony of the cocktail waitresses, the one bartender who testified, and one of the beverage supervisors employed at Resorts in 1978, all indicated that anywhere from 60 to 90-percent of the alcoholic drinks served in 1978 contained at least 2 ounces or more of liquor.
Another significant area of unreliability in respondent's calculations, as we explained previously herein, is the use of a total figure of 51,719 for the total hours worked on the casino floor by the cocktail waitresses. In light of the unreliability of that figure, we would prefer not to rely on its accuracy.
Results of the IRS Surveillance
The reservations we have as to the figures used in respondent's calculations for the number of total drinks served on the casino floor and for the total hours worked by the cocktail waitresses on the casino floor, and the reservations we have as to the indirect method*491 used by respondent to calculate his estimate of the average tip rate per hour of the waitresses, cannot be ignored in light of the results of respondent's own surveillance which provides statistically reliable data from which a more direct and more accurate estimate of the actual tip rate per hour can be calculated.
Prior to explaining our use of respondent's surveillance data to avoid the unreliable features of respondent's calculations of the total number of drinks served and the total hours worked on the casino floor in 1978 by the cocktail waitresses, we will comment on a number of questions raised by petitioners concerning the surveillance.
As we noted earlier, respondent used the surveillance data only to calculate the tip rate per drink and calculated a rate of $ .81 per drink. Respondent made no adjustment for the 95-percent or 97 1/2-percent lower confidence limits concerning which both experts testified. In Catalano v. Commissioner,supra, this Court approved the use of the figure reflecting a 95-percent lower confidence limit (81 T.C. at 16), and we do so here, adopting an average tip rate per drink of $ .77.
Petitioners object*492 to the use of a tip per drink ratio (or other statistics) based on a surveillance conducted in 1980 and 1981, in order to calculate and reconstruct tip income for 1978. The basis of their objection is that the conditions prevalent at Resorts in 1978, particularly due to the large crowds, were so different that use of information obtained from a surveillance conducted in later years is unreasonable. However, petitioners have produced no studies or other evidence that would convince us that people on the average leave lower tips under crowded conditions than they do under uncrowded conditions. Although it may be more difficult for waitresses to serve drinks quickly when there are crowds in the casino, the presence of more customers (and therefore more potential tips) in 1978 may, on the average, have resulted in more tips being paid than in uncrowded conditions.
Petitioners also make various attacks on the validity of the surveillance study itself, contending that respondent's agents could not possibly have seen the actual exchanges of tips well enough to observe them accurately, that they could not mentally have kept an accurate running account for up to half-an- hour of tips received, *493 and that insufficient instructions were given to enable the agents to distinguish non-valued chips (which are used by players at the roulette wheels, each of whom receives his or her own identifying color) from valued chips.
Similar objections concerning respondent's surveillance procedures were raised and rejected in the Catalano case, and we reject them here as well. The agents involved herein testified that they had little or no trouble making the observations, remembering the necessary numbers or otherwise carrying out their assignments. 7 Respondent's statistical expert subjected the results of the surveillance to a number of statistical tests including various ratio analysis calculations to measure the reliability of the surveillance results and such tests supported the validity thereof. We accept his testimony that the data collected exhibited a low degree of "variability," and that the "variability" would have been much higher if the agents had had difficulty making the observations or otherwise following instructions. Even petitioners' expert agreed that the surveillance data exhibited "little variation".
*494 Court's Approach
We adopt an approach that bypasses the two significant areas of unreliability in respondent's calculations (the total drinks served and the total hours worked on the casino floor), and we rely primarily on the results of the surveillance, which results have been found to be statistically valid.
Because the 1981 surveillance involved observations of drinks served and tips received during a large, randomly selected sample of discrete half-hour intervals over a 6-month period of time, we believe that that surveillance data can be used to calculate a drink per hour rate that is as reliable as the tip per drink rate. With a tip per drink rate and a drink per hour rate (both derived from the I.R.S. Surveillance data), only a simple arithmetic calculation is required to obtain an average tip per hour rate which can be applied to the hours worked of each cocktail waitress on the casino floor.
The surveillance data for 1981 reflects that 537 drinks were served over 59 half-hour observation periods. Dividing 537 drinks by 29.5 hours equals 18.2 drinks per hour. Utilizing the tip per drink ratio as previously discussed, we calculate the gross average hourly*495 tip rate as follows: $ .77 tip per drink times 18.2 drinks per hour equals $14.01 in tips per hour.
Respondent's calculations do not allow for the fact that the cocktail waitresses usually shared a portion of their tips with the bartender and the barboy to assure that the cocktail waitresses received prompt and accurate service in the future. Such adjustments previously have been allowed where the facts indicated that they were appropriate. 8 Indeed, respondent made adjustments for that practice in calculating average hourly tip rates of employees at other locations within Resorts that were part of the same tip income project, but not for the waitresses working on the casino floor. The uncontested testimony of petitioners herein was that generally they paid $1 to $5 of their tips per shift to the bartender and usually $ .50 per shift to the barboy. However, tipping was "according to service" and did not necessarily occur every shift. The only bartender who testified indicated that he generally received $2 to $4 per shift from each cocktail waitress who worked on the casino floor.
*496 Considering the record as a whole, we find that, on the average petitioners split $2.50 of their tips per shift with the service bartenders and barboys. Therefore, we make a $2.50 per six-hour shift downward adjustment in the tip rate. This adjustment converts to a $ .42 downward adjustment per hour, producing an average tip rate per hour of $13.59.
Finally, we recognize that petitioners herein were not available to receive tips during their lunch time. Petitioners agreed that they were allowed one 30-minute lunch break per six-hour shift, and they testified that they took additional breaks of either five minutes per hour or five minutes per half hour. However, the break time each hour during which no tips could have been received was reflected in the I.R.S. surveillance figures. The 30-minute lunch hour break was not reflected because a substitute or "breaker" waitress would then take the place of the regular waitress. Therefore, petitioners are allowed a total of 30 minutes or one-twelfth of each shift as break time. Multiplying the previously adjusted average hourly tip rate of $13.59 by 11/12 (approximately .917) results in a net average hourly tip rate of $12.46 per*497 hour. 9 We hold that $12.46 is the hourly tip rate that should be used in reconstructing petitioners' tip income received during 1978 for serving complimentary drinks on the casino floor of Resorts. Since that figure already accounts for average break time, the average tip rate should be applied in full to each hour of each shift to which each waitress was assigned to work on Resorts' casino floor.
Section 6653(a) addition to tax
Respondent has determined that each of the petitioners herein is liable for the 5-percent addition to tax for negligence or intentional disregard of rules and regulations. The burden of proof is on petitioners to show that the additions were improperly imposed. Enoch v. Commissioner,57 T.C. 781, 802 (1972).*498 We uphold respondent on this issue as to each of the petitioners herein.
Several of the petitioners herein (namely, Margo Achille and Cathy Stutzman) were college graduates in 1978. Several others (namely, Clara Fesco and Bonnie Gasperini) had prior experience as waitresses. Theresa Bruno testified that she kept a tip income diary in 1978, but did not retain it. Cathy Stutzman submitted a "diary" for 1978, but it was prepared in 1981 and was incomplete. For example, it contained no entries for December, although she did work that month. All petitioners reported at least some tip income, except for petitioner Achille who testified that she understood that the W-2 statement received from Resorts and showing $1,995 in income included all of her tips. All petitioners herein indicated on the witness stand that they understood their legal obligations to report all of their tip income.
In light of the above facts, the failure of each petitioner herein to maintain and keep contemporaneous diaries or other records of actual tip income received in 1978 constitutes either negligence or intentional disregard of the rules and regulations contained in section 6001. We therefore hold that*499 each petitioner is liable for the section 6653(a) addition to tax.
Appendix
Because of the disputed nature of the number of hours worked and/or the locations within Resorts at which the hours were worked by the individual petitioners herein, we make the following findings of fact with regard to each petitioner:
Achille, Margo A. - Docket No. 13439-82
Respondent has determined that Margo Achille worked 629 hours gross on the casino floor. Petitioner does not dispute this figure, except to contend that it should be reduced to 509 net hours to allow for breaktime and prep time. Because the average hourly tip rate that we have determined reflects an adjustment for break time (and because the evidence indicated that the waitresses were not paid for prep time prior to the beginning of each shift), we find the 629 hours to be accurate. Therefore, petitioner Achille's tip income received at Resorts in 1978 is to be computed as follows: 629 hrs. times $12.46 per hour equals $7,837.34 total tip income.
Bruno, Theresa J. - Docket No. 9578-82
The records obtained from Resorts indicate that Theresa Bruno worked 463 hours on the casino floor and 211 hours in The*500 Gold Room, for a total of 674 gross hours. Petitioner's records, as derived from her time cards, indicate a total of 601.5 hours gross and 498 net hours for the two locations.
Because of the demonstrated errors in Resorts' pay proofs from which respondent derived his figures, and in light of Ms. Bruno's testimony, we accept her figure of 601.5 gross hours. However, nowhere does Ms. Bruno indicate how those hours break down between the two locations at which she worked, (namely, the casino floor and the Gold Room). In the absence of any other evidence, we therefore assume that the ratio of hours represented by respondent's figures is correct - that is, 31.3-percent of her hours were spent working in the Gold Room and 68.7-percent were spent working on the casino floor. Accordingly her tip income is reconstructed as follows:
| Location | Hours | Hourly Rate | Total Tips |
| Casino Floor | 68.7% X 601 = 413 10 | $12.46 | $5,145.98 |
| Gold Room | 31.3% X 601 = 188 | 3.92 | 736.96 |
| Total Tip Income | $5,882.94 |
Fesco, Clara M. (a.k.a. Marlisle Fesco) -*501 Docket No. 27315-82
There are a number of decrepancies in the record concerning the hours Ms. Fesco worked. Respondent calculated 777.5 hours, but the breakdown as to time spent at various locations is unclear. Petitioner submitted a reconstruction based upon her time cards of hours worked in 1978. Her construction reflected total hours of 533. However, her reconstruction is on a net basis, and her methodology in determining those net hours is not entirely clear. It is apparent that she deducted at least 60 minutes for break time per shift, but sometimes she would deduct more.Her testimony indicated that she subtracted the last 15 minutes of every shift as part of prep time, but her method of recollecting and/or reconstructing the rest of her prep time is not clear or consistent.
Ms. Fesco testified that her time cards indicate a total gross number of hours worked of 715 hours. 11 She also testified that she worked twelve days in the Rendezvous Lounge, ten days in the Gallery, and the balance on the casino floor. We found petitioner's testimony to be credible. An examination of her reconstruction indicates that she worked 72 gross hours in the Rendezvous Lounge and 63*502 gross hours in the Gallery, and we so find. The balance, or 580 hours, must therefore have been worked on the casino floor, and we so find.
Ms. Fesco's 1978 tip income received at Resorts is reconstructed as follows:
| Location | Hours | Hourly Rate | Total Tips | ||
| Casino Floor | 580 | X | $12.46 | = | $7,226.80 |
| Rendezvous Lounge | 72 | X | 7.12 | = | 512.64 |
| Gallery | 63 | X | 5.56 | 350.28 | |
| Total Tip Income | $8,089.72 |
Gasperini, Bonnie - Docket No. 13484-82
Respondent's calculations with respect to Bonnie Gasperini were based on Resorts' payroll records and reflect 297 hours in the Rendezvous Lounge, 383 hours on the casino floor, and 288.5 hours in the Gold Room. Petitioner testified that she never worked in the Gold Room, but split her time between the Rendezvous Lounge and the casino floor. The schedule of hours she submitted, based upon her*503 time cards, showed totals only for the net hours. Our examination of her time cards reflects a total of 321 1/3 gross hours on the casino floor and 485 1/6 gross hours in the Renezvous Lounge, and we so find.
Therefore, Ms. Gasperini's reconstructed 1978 tip income received at Resorts is as follows:
| Location | Hours | Hourly Rate | Total Tips | ||
| Casino Floor | 320 | X | $12.46 | = | $3,987.20 |
| Rendezvous Lounge | 485 | X | $7.12 | = | 3,453.20 |
| Total Tip Income | $7,440.40 |
Stutzman, Cathy A. - Docket No. 11000-82
Respondent, relying on Resorts' records, reconstructed Cathy Stutzman's tip income in part on the basis of 134 hours worked in the Rendezvous Lounge, and 158.5 hours worked in the Gold Room. We accept Ms. Stutzman's testimony that she did not work at either of those locations in 1978, but rather worked all of her hours on the casino floor.
The reconstructed schedule of hours worked submitted by Ms. Stutzman, based on her time cards, reflects a total of 446 net hours. An examination of those records indicate that she subtracted only 60 minutes of break time per shift (in addition to "prep time" prior to the shift). Therefore, she*504 would generally record "5" hours worked out of a shift of 6 hours.
We therefore multiply her net hours by 6/5 or 1.2, and find that she worked a total of 535 gross hours, which when multiplied by the hourly tip rate for the casino floor of $12.46 per hour, results in total tip income of $6,666.10 received at Resorts in 1978.
In accordance with the findings and holdings herein,
Decisions will be entered under Rule 155.
Footnotes
1. Cases of the following petitioners were consolidated for trial, briefing and opinion: Theresa J. Bruno, Docket No. 9578-82; Cathy A. Stutzman, Docket No. 11000-82; Margo A. Achille, Docket No. 13439-82; Bonnie Gasperini, Docket No. 13484-82; Clara M. Fesco (a.k.a. Marlisle Fesco), Docket No. 27315-82.↩
2. All section references are to the Internal Revenue Code of 1954, as in effect during the year in issue.↩
3. Complimentary drinks were not served in the slot machine area in 1978.↩
4. In 1978, the minimum wage was $2.65 per hour. Petitioners evidently were paid $2.29 per hour by Resorts, on the assumption that they would make at least $ .36 per hour in tips. The wages paid of $2.29 per hour and the assumed tips of $ .36 per hour were treated by Resorts as wages subject to tax withholding and were reported on petitioner's tax returns as "wages."↩
5. A lower "confidence limit test" refers to the use by a statistician of a number lower than the number produced by a sample. The statistician can then state that if the sample was repeated he is certain (expressed in percentages), that the average that would be obtained in the new sample would not be below the average obtained in the first sample after reduction by the appropriate confidence limit test. See, e.g., W. Cochran, Sampling Techniques, sec. 2.8, p. 27 (3d ed. 1977).↩
6. Court-initiated adjustments to respondent's reconstructions of income previously have been made by this Court in other tip income cases where the factual circumstances of the case warranted the adjustments. See, e.g., Brancaleone v. Commissioner,T.C. Memo. 1963-318, Yukimura v. Commissioner,T.C. Memo. 1982-58 and Francis v. Commissioner,T.C. Memo. 1985-6↩.
7. The few discrepancies in the agents' testimony concerning the details of the surveillance procedures were minor.↩
8. See, e.g. Feldman v. Commissioner,T.C. Memo. 1983-781 and McQuatters v. Commissioner,T.C. Memo. 1973-240↩.
9. We make no specific allowance and conclude that there should be none for the last fifteen minutes of a shift during which there was some testimony that the waitresses cleaned up, did not serve drinks and did not receive tips. Such time periods were included in the surveillance data, and have therefore already been taken into account.↩
10. Throughout the calculations in the appendix, we have rounded to the nearest whole hour where necessary.↩
11. The discrepancy between respondent's figure of 777.5 hours and petitioner's figure of 715 hours may be explained in part by Resorts' policy of paying an employee for an entire 6-hour shift, even though he or she was sent home early due to lack of business in a particular location of the casino.↩