Critical Analysis of the report, Economic Impact: Tomatoes in Florida, Report 1

produced by the Center for Reflection, Education and Action

Bruce Nissen
Director
Research Institute on Social and Economic Policy (RISEP), Florida International University
nissenb@fiu.edu
305-348-2616

The report Economic Impact: Tomatoes in Florida, Report 1 was commissioned by the McDonald’s Corporation and completed by the Center for Reflection, Education and Action (CREA) in April 2006.  An examination of the conditions and pay of tomato workers in Immokalee, Florida is important and useful, provided the report has scientific validity and objectively studies the circumstances of these workers. 

Unfortunately, the work is so riddled with errors both large and small that it cannot be accepted as factually accurate on virtually any measure.  The report fails to explain key questions (both methodological and substantive), contains numerous errors including simple matters of arithmetic, leaves out crucial pieces of information needed to understand the meaning of its supposed “findings,” and arrives at conclusions that are based on false or non-existent comparisons and issues of judgment having no relationship with the study or its attempt at supplying data. 

Therefore, little of substance can be concluded from the report, and to the extent that conclusions can be drawn, they would contradict what appears to be the author’s main intent:  to argue that workers picking tomatoes for McDonald’s tomato suppliers are already better off than they would be if McDonalds agreed to the demand by the (CIW) to pay growers 1 cent more per pound of tomatoes with that cent passed on to the workers in wages.  In fact, even taking the basic figures obtained from one McDonald’s grower at face value, the opposite conclusion would be required.

Wage figures in the report

We start with what should have been the biggest red flag to those putting out the report.  Utilizing the records of a grower that supplies McDonalds presented as “Grower #1” for the year 2005, the report states that workers are paid 45 cents per 32-pound bucket.  Chart 1 (page 12) claims that at that rate of payment per bucket, the 10 fastest pickers for this grower earned an hourly wage of $18.27 per hour, while the 10 slowest pickers earned $9.65 per hour in 2005.  Chart 2 (page 13) looks at 7 crews and determines that average hourly pay in the crews ranged between $11.88 per hour and $17.74 per hour. 

Without looking any further, these figures are breathtaking and entirely out of line with anything we know about farmworkers’ wages.  According to state and federal figures, the wages for “farmworkers and laborers, crop, nursery and greenhouse workers” in 2005 in the state of Florida were as follows: 

Median:  $7.26 per hour
Mean:  $7.61 per hour
Entry level:  $6.15 per hour
Experienced level:  $8.26 per hour
(Data from the Florida labor market website: www.labormarketinfo.com/library/oes.htm )

If the median in the state is $7.26 per hour, how could these workers be earning such fantastically higher wages than anybody else doing similar work?  At the very least, employer reports of wages 64% to 144% above the norm should raise a few eyebrows and invite a closer investigation.  Yet the CREA report never mentions government data, and seems to be unaware of how incongruent its reports of wages are with all other data available. 

However, a closer analysis of the data in the report only increases one’s unease about its reliability.  Simple elementary tasks such as multiplication and division seem to be done wrongly.  Chart 1 (page 12) incorrectly multiplies the number of buckets picked by pay rate (X cents per bucket) and reports wage rates incorrectly for every single rate of pay at both speeds of picking.  The hourly wage rates are only off a few cents for every rate of pay (sometimes too high by as much as 22 cents, sometimes too low by as much as 20 cents).  Perhaps deviations this small can be explained because numbers are rounded in the Chart and are not in the computation of the pay rates.  In any case, simple multiplication of the numbers in the chart arrives at a wage that deviates from the wage rates given in the chart.

Unfortunately, the errors only grow worse from there.  Out of eight hourly wage rates given in Chart 2 (page 13), six are incorrect, and two are impossible to judge because sloppy reporting of the data used makes it impossible to check the wage rate stated.  Five of these crews are reported to be picking regular tomatoes at 45 cents per bucket.  All five have incorrectly computed wage rates, which can be obtained simply by multiplying the number of buckets picked by 45 cents.  (Again, perhaps rounding can explain some of the deviations, but for Crew #1, the deviation is $2.71 per hour – well beyond anything that rounding errors could explain.)  One crew is reported to be picking Roma tomatoes, which is paid at 65 cents per bucket.  It too reports an incorrect wage rate, which can be obtained by simply multiplying the number of buckets by 65 cents.  Two of the crews are reported to pick both regular and Roma tomatoes, and since no information is given concerning how many of the total buckets picked are of each type, it is impossible to check the wage rates given.  (It should be elementary research procedure to specify the “mix” of products picked so that results can be retabulated and verified.)  But for all six where results can be checked, the final result is wrong by amounts ranging from trivial (rounding error?) to large. 

But the data errors don’t stop there.  An unnumbered chart (it would be the third chart in the report) on page 15 lists data taken from Grower 1’s January and February 2006 records.  The ostensible purpose is to determine if a new formula being attempted by this grower results in superior pay for the pickers to the old formula of 45 cents per bucket.  The new formula is the state’s minimum wage ($6.40 per hour) plus 10 cents more per bucket picked.  The data presented for these 10 days and the accompanying explanation are particularly opaque, making it possible to interpret data a number of ways.  But no matter what method one uses to analyze the data, the “average hourly wage for the day” presented in the final column is impossible. 

The numbers in the chart are explained very obscurely.  It is noted that “The data received was for a workday that had been limited to 6 hours per day as a result of the damage from the hurricane and the freezes that hit the area in prior months” (p. 14).  But, wanting to see what the impact of the new wage formula would be in more normal times with a normal 8 hour day, “we have pro-rated these numbers to see what the wages earned would be at the same picking rate for the workers for an 8 hour work day” (p. 14).  So far, so good:  Thus it would seem that the “Average number of buckets picked” in the table refers the number that would be picked in 8 hours.  Also, it would seem that the “Average wage earned for the day” would refer to earnings in 8 hours, and the “Average hourly wage for the day” would refer to an 8-hour day.  If we take this straight-forward reading of the chart, the numbers presented are way out of line. 

Just to take a somewhat “typical” day, the figures for “Day A” can be used.  Ignoring for the moment the “lowest” and “highest” columns, the chart shows that in Day A the average # of buckets picked was 121, the average wage was $63.33, and the average hourly wage for the day was $10.56.  Simple arithmetic shows that the average hourly wage is wrong:  $63.33 divided by 8 hours equals approximately $7.92 per hour, not $10.56 per hour.  The author of the report has apparently divided daily pay by 6, not 8, and therefore has come up with artificially inflated wages usually over $10 per hour. 

Perhaps because of the unclear way the data are explained, there is some other meaning to these data, although multiple ways to interpret them fail to make sense of the report’s conclusion that high wages often exceeding $10 per hour will result.  For example, converting the average data into a “cents per bucket” payment (i.e., dividing the daily wage by the daily # of buckets picked) reveals that pickers were paid as low as 41 cents a bucket (Day F), 44 cents a bucket (Day J), up to $3.30 per bucket (Day D), depending on rate of picking. 

The one clear thing that comes from the data is that the payment per bucket increases for the less efficient and decreases for the more efficient under the new formula.  Days with high pickage rates (Days F and J, for example) result in payment lower than the old rate of 45 cents per bucket.  Days with low pickage rates (the extreme example being Day D), result in higher payment per bucket.  But for the fastest-working pickers, the new formula results in pay decreases, some of them fairly substantial.  For example, the chart shows that in Day B, the fastest picker picked 200 buckets, for a pay of $71.20.  That translates into less than 36 cents per bucket, well below the 45 cents per bucket of the older formula.

Another way to check the average hourly rates claimed in this chart is to determine how many buckets per hour a worker would have to pick under the minimum wage plus 10 cents a bucket formula.  Again, using Day A as a fairly typical day, this exercise indicates the fantastically high rate of productivity that would be necessary to earn the wage claimed as the average hourly wage.  To earn the $10.56 per hour this chart claims pickers earned in Day A, a picker would have to pick approximately 42 buckets per hour – over 1/3 of the entire number of buckets picked in the entire day, according to the figures provided! (see footnote at end of this report for an explanation of this calculation by clicking here)

Yet the report’s own figures in previous charts show that the average number of buckets picked per hour is well below 42.  Chart 2 (p. 13) shows that in 2005 the average buckets picked by different crews ranged from 21 per hour up to 39 per hour.  One hesitates to give these data too much credence, given the numerous discrepancies already uncovered, but even accepting them at face value, practically nobody is able to produce 42 buckets per hour.  Thus, the $10.56 per hour wage is a fantasy. 

By the time one is finished looking at these charts, virtually nothing in them concerning average pay holds up.  Recalculating the average hourly pay in the final column of the daily chart on page 15 by using an 8 hour day (i.e., by dividing by 8 rather than 6), results in the following adjustments: 

Day                 Average hourly pay claimed                        Average hourly pay (actual)
A                                 $10.56                                                 $7.92
B                                 $10.71                                                 $8.03
C                                 $9.87                                                   $7.40
D                                 $8.80                                                   $6.60
E                                  $10.42                                                $7.82
F                                  $11.29                                                 $8.47
G                                 $10.87                                                 $8.15
H                                 $10.29                                                 $7.72
I                                   $10.76                                                 $8.07
J                                  $11.07                                                 $8.30

It goes without saying that the bolded wage rates in the final column are much closer to government figures of wages for farmworkers in Florida.  The report being critiqued simply has no credibility in its reportage of wage rates at Grower 1.  

Additional problems with the report

The report makes another simple ground-floor error when calculating the value of employer-provided housing (page 19).  It argues that outside housing would cost between $25 and $40 per week (average $32.50 per week).  The report states that “there is available to the workers living in the Grower provided housing the equivalent of this $32.50” but fails to subtract the $14 per week being paid to the grower for this housing.  Thus, the housing subsidy from the grower, if we accept all the figures as given, is not $32.50 per week, but $32.50 – $14.00, or $18.50.  The $18.50 per week subsidy amounts to $2.64 per day.  This is not an insignificant sum for such low-paid workers, but it is also nowhere near the equivalent of the almost $1 per hour increase that an extra $32.50 per week (for those working a 40-hour work week) would be. 

There are additional errors in the report indicating the author’s lack of knowledge about Florida and its low-wage workers.  The report states, “Florida’s first minimum wage became effective on May 2, 2005 when the Florida State Legislature set the state minimum wage at $6.15 per hour” (page 12).  This is false.  The majority of the Florida state legislature was and is opposed to a state minimum wage in Florida, as is the governor.  The new minimum wage came about as a result of a ballot initiative amending the state’s constitution, a drive spearheaded by Florida ACORN (Association of Community Organizations for Reform Now) and opposed strenuously by the state’s leading politicians. 

Lack of knowledge of Florida farmworker payment systems is also demonstrated by the report’s presentation of the “new” formula of Grower 1 to pay workers the minimum wage plus 10 cents a bucket, rather than 45 cents per bucket.  This formula is presented as (a) new; (b) undertaken by the grower only to see if it could “be more beneficial for the workers” (p. 14); and (c) apparently beneficial for the average worker (page 14).  These are either outright false or highly dubious claims. 

First, the “day rate (hourly) plus a dime per bucket” system is a very old system in Florida farm labor relations.  A strike in 1995 at Pacific Land Company was over the company’s attempt to cut the “day” portion of the “day plus a dime” system to a rate below the federal minimum wage.  This system has been around for a long time.  Second, the notion that the grower is experimenting with this “new” system solely to see if it could be beneficial to the workers is a novel way to interpret employer motives, but few private market employers have a sole motive of worker welfare improvement (after all, a simpler way would simply to be to raise the workers’ wages, if the only motive is helping them out).  Instead, the most likely motive (consistent with its status as a for-profit employer) is that the employer is experimenting to see if it can make a higher profit from the “new” system than it did under the old one. 

Finally, the implied conclusion (page 14) that the new system is more beneficial for the workers (meaning it pays them more) is brought into question by the report’s own figures.  As an illustration, consider the first three days (Days A, B, and C).  Under the “day and a dime” method of payment, these workers picked 121, 131, and 80 buckets.  They were paid $63.33, $64.27, and $59.20 for those three days.  These are fairly typical days, with one of them having an exceptionally low number of buckets picked.  Compare what the workers would earn under the two systems of payment (“day and a dime”, or “45 cents per bucket”). 

Day     Day and dime method of payment            45 cents per bucket method of payment
A                     $63.33                                     $67.25      (2 hrs. X $6.40 + [.45 X 121])
B                     $64.27                                     $71.75      (2 hrs. X $6.40 + [.45 X 131])
C                     $59.20                                     $48.80      (2 hrs. X $6.40 + [.45 X 80])

From the above simple calculations, it is apparent that on two of the three days, the workers would be better off under the old system paying 45 cents per bucket (by $3.92 on Day A and $7.48 on Day B).  On Day C, on the other hand, the “day and a dime” system pays considerably better (by $10.40). 

The conclusion one could draw from this is that the “new” system is beneficial to the very slowest pickers, and on days where weather and other conditions keep rates of picking abnormally low.  But for faster pickers, and on more “normal” days, the old system of 45 cents per bucket is preferable.  The bigger question is which of the two systems maximizes the workers’ wage bill, and the data given indicates that the old system does that about 70% of the time!  Contrary to the report’s implication that the “new” system is “more beneficial to the workers,” in most cases the opposite appears to be the case, although the new system may maximize the grower’s profits. 

A deeper appreciation of the dynamics of pay systems and worker motivation would indicate that nothing definitive can be deduced from the above figures.  Perhaps under one system the workers will exert more effort than under the other.  While it would take us far afield, a precise conclusion about the impact of different pay systems on the workers’ pay would require more information than we have at present, although the preliminary data seem to show the “new” system as actually detrimental to the workers, contrary to the report’s apparent conclusion. 

Numerous other methodological issues and questions could be raised about this report.  Almost nowhere are ordinary norms of social science research followed.  Necessary information to judge data and conclusions is frequently missing, making it difficult to evaluate anything.  Any protocol used, or set of survey questions, or circumstance of worker surveys is largely missing.  Results from the surveys (or interviews) of workers are presented so generally that we don’t know ranges, percentages, numbers, or anything else.  Sample sizes are set for crews by a method nowhere explained, and given the great diversity in crew size, such details matter.  Much of the same is true for the housing data and results.  These types of problems make it difficult to evaluate much of the rest of the “results” presented in the report.  Utilizing just one grower to represent all McDonald’s suppliers is of course very problematic.  So is accepting at face value all data and records supplied by that grower.   But we will leave these problems aside, and proceed to what appears to one of the primary purposes of the report:  to prove that McDonald’s suppliers are paying their farm laborers more than would be the case if they agreed to the demand by the (CIW) to pay a “penny a pound” more to the workers. 

The report concludes (p. 21), that “the wages and benefits provided by Grower 1 exceed the wages earnable at 40 cents per bucket plus a ‘penny a pound’.”  For this conclusion to be true, the wages and benefits would have to exceed 72 cents per bucket, since an average bucket weighs 32 pounds.  Do the data in this report show this? 

One hesitates to use any data from a report so riddled with errors, but if we try to use the data within it that have likely had the least amount of potentially contaminating intervention (as well as the one leading to corrected wage figures that are closest to what we know from government sources about farm worker wages), we probably should use the raw figure for “average number of buckets picked” to test the claim.  Only in one chart (p. 15) are raw figures for buckets picked given, for 10 days chosen at random in January and February of 2006.  Adding those figures over the 10 day period (1156 baskets) and dividing by 10 shows an average of 115.6 buckets per day.  At this “average” rate of picking, is it true that Grower 1 is paying the workers more than 72 cents per bucket? 

115.6 buckets times 72 cents would provide a wage of $83.23 per worker per day.  In contrast, Grower 1 paid an average of $62.77 per day.  Even adding a “housing subsidy” of $18.50 per week or $3.70 per work day in a five day work week ($2.64 per day in the 7 day week) that some of Grower 1’s workers have, we only arrive at a wage of $66.47.  That is a shortfall of $16.76 per day, or $83.80 per week assuming a 5 day work week or $100.56 assuming a six day work week.  The facts completely contradict the report’s claim that “the wages and benefits provided by Grower 1 exceed the wages earnable at 40 cents per bucket plus a ‘penny a pound’.” 

Of course we must be careful relying on any “facts” contained in this report, but the above analysis is the closest we could get to daily numbers on buckets picked per day, which is the only way to compare the present wages of Grower 1’s workforce, and the wages that would be available if a penny a pound was added to 40 cents per bucket. 

Nowhere in the report is there an analysis of what the wage impacts of a “penny a pound” increase would be.  For good reason:  such an analysis would show that wages would be raised 91% for those workers currently being paid 35 cents a bucket; 80% for those currently being paid 40 cents and bucket; and 71% for those currently earning 45 cents a bucket. Wage increases of that magnitude are so large that the proposition that an individual grower such as Grower 1 could pay his workers at a higher rate is simply implausible.  It would be very difficult, if not impossible, for a grower to fund such a pay rate and remain competitive.  That is why the is calling on the corporations that purchase tomatoes to fund the increase through a surcharge that would be passed through the grower directly to the workers to increase their per bucket pay.

This analysis has not focused on the other parts of the report under examination (such as the S.A.F.E. program, McDonald’s Grower’s Standards, the issue of full-time year-round work, and the like).  Rather, it has focused on the fantastic wage claims made in the report, and the attempt to discredit the call by the (CIW) for large pay increases through a “penny a pound” pass-through to worker wages.  Regarding those issues, the report should have no credibility whatsoever. 

 

The report under consideration was accessed on April 25, 2006 at this web site:
http://www.crea-inc.org/tomatoproject.htm .  

The author of this critique has not been paid by anyone to produce it.  The author is the Director of Research at the Center for Labor Research and Studies, Florida International University, as well as the Director of RISEP. 

The calculation is rather straightforward:  Working an 8 hour day, a worker would earn from the minimum wage $6.40 X 8 hours, or $51.20 daily without the 10 cents per bucket bonus.  Working at $10.56 per hour, the worker would earn $10.56 X 8 hours, or $84.48 daily.  The difference ($84.48 – $51.20) is $33.28.  $33.28 over an 8 hour day comes to $4.16 per hour.  To earn an extra $4.16 per hour, at 10 cents a bucket, a worker would have to pick 41.6, or almost 42 buckets per hour.  If the worker is picking for only 6 of those 8 hours (doing other fieldwork the other two), the required number of buckets climbs even higher, to 55 buckets per hour – a pace that no one claims any worker can do.