Thursday, June 2, 2016

The Devil is in the Definitions

We put a great deal of emphasis on quantifying things. We classify, define, count and report. It is a natural outgrowth of the modern era, the availability of data, the increasingly availability and decreasing cost of computers and software. Gathering and interpreting data has become simpler and cheaper. And society seems to love and even crave it. The OJCC began getting questions about the volume of filings recently, with various folks suddenly interested in an update of their data. 

Albert Einstein and William Cannon are each credited with saying "not everything that counts can be counted, and not everything that can be counted counts." It is a poetic play on words and is actually quite enlightening. At the root of this truism, and at the root of data analysis generally, is the fundamental requirement of definition. One cannot begin to quantify until one defines. That is, what is it that will be counted, and how are those things defined? And the "devil may be in the definitions."

We have seen this result in debate and controversy in various analyses. Since 1930 there have been nine planets in our solar system. More recently, someone thought they discovered another, beyond Pluto, which was named Eris, and then there were ten. That led to discussion of the critical point, "what is a planet." And the result was not focused on whether Eris is or is not, not whether there are nine or ten, but on whether either Eris or Pluto is or is not and actually whether there are eight or ten. Pluto lost. 

Another example in 2010, there was a great deal of debate about unemployment compensation. According to the USA Today, by 2010 "so many Americans have been jobless for so long that the government is changing how it records long-term unemployment." The Feds perceived "an unprecedented rise" in long-term unemployment, and so the federal Bureau of Labor Statistics (BLS)," changed its data parameters. It increased "the upper limit on how long someone can be listed as having been jobless" from two years to five years. By increasing the parameter, the definition of "unemployment" it necessarily increased the count. 

If we are counting "frequent beach visitors" and our parameter is "have you been to the beach in the last two years," that will yield a quantity of positive responses. If we instead ask "have you been to the beach in the last five years," that will yield a different quantity. Since the responses to the second question (5 years) necessarily include the responses to the first (2 years), the quantity yielded by the more expansive, second, question can be expected to be larger. Although it is possible that it would be the same, that is frankly statistically unlikely. By defining what is "frequent" the volume of responses can potentially be manipulated.

Thus, we understand that the definitions used will likely have an influence on the outcome of our statistical analysis. Whether we are counting planets, the "long-term unemployment," frequent beach visitors, or even "work" injuries. We have to remain cognizant of a primary rule of data analysis is a computer term "GIGO," which stands for "garbage in, garbage out." I am not sure who coined the phrase, but essentially it is accepted as meaning that flawed input will result in flawed output when analyzing data. 

To test the acceptance of GIGO, I recently asked 25 people two questions: "are you a computer expert," and "do you agree that appropriate output is dependent on the quality of input." ninety-six percent of the respondents described themselves as "computer experts," and all of those agreed with the GIGO concept. Some might suggest on this strong empirical, statistical evidence, the GIGO concept is immutable. 

The recent 2016 Workers' Compensation Summit provided an opportunity to hear a vast array of perspectives from people involved in American workers' compensation. One of the topics raised was federalization of workers' compensation. There is a belief that the federal government will ultimately interfere with workers' compensation, insinuating itself at least in the context of how workers' compensation interacts with the Social Security system. We have seen this model already in the interference and bureaucracy of the Medicare system. Anyone that has studied that interaction well knows the perils of federal involvement. Others believe that more dramatic federal invasion is inevitable. This is an intriguing discussion. 

After the Summit, I ran across an article summarizing conclusions reached recently by John Burton and Steve Guo. Professor Burton is an academic who has extensively studied workers' compensation, and was Chair of the 1972 Commission. There are those who worship at the altar of the 1972 Commission Report, and revere its conclusions and recommendations. There is also a population that finds many faults with both the process and outcome of that Commission. I find most compelling the adherents and critics who can intelligently describe their position or conclusions with logic and reason. I find I have little time for those who either hate or love the 1972 Commission report but struggle to explain their position. 

Several of the conclusions outlined in this recent article summarizing Burton's recent work are worthy of discussion. 

One example is that Burton and Guo reach conclusions about the extent of claims that are in the SSDI system, which allegedly should also be in a state workers' compensation system. They point to "heightened Standards of Proof," in a variety of states, Florida among them. The contention is that the "major contributing cause" standard prevents payment of workers' compensation benefits in the modern era. This is compared to a time when they contend workers' compensation would provide benefits if the work environment contributed to the injury in any manner beyond "trivial." 

There is a contention that "cash benefits," more aptly wage-replacement benefits, "have trended negatively." The authors contend that the few available studies indicate that some decrease is due to fewer injuries (decreased "frequency"). The larger share of decrease, they contend, is some combination of three factors: (1) state systems are more stringent, (2) eligibility under state acts has been tightened, and (3) fewer individuals qualify for permanent partial disability benefits.

The authors contend that "practical or actual replacement" amounts for lost wages have been demonstrated to be well below the "roughly" 66% parameter common in workers' compensation laws. They define this as a "shortfall" and contend that it is this shortfall that may drive injured workers to the SSDI system. This analysis, however, focuses on the replacement "10 years after injury." Is this a failure in the calculation of benefits? Some might argue it is a failure to account for the effects of inflation instead.

Another contention is that SSDI is being saddled with costs related to work-place injuries. The authors note that if a person receives SSDI and some form of state wage-replacement, and the two exceed a formulaic maximum under federal law, then the SSDI payment is reduced (called an SSDI Offset). This is essentially an arbitrary legislative (federal) determination that some level of income while disabled is "enough." They also note that some 15 states have reverse-offset provisions, in which the state benefit is instead reduced, contending that this burdens the federal system. 

The authors concede that the "SSDI pays benefits regardless of the source of disability." It is worthy of note that Social Security is at its root a socialistic program into which all working Americans pay. Most wage earners notice that 7.5% tax (FICA) that is taken from their paycheck to fund Social Security. Most likely do not notice that there is an additional 7.5% tax that is likewise contributed by their employer in "matching funds." Essentially, every American wage earner contributes 15% of their income to support the Social Security system. One could argue whether that second "matching fund" percentage would become pay or company profit in an alternative, but the fact remains that 15% is contributed now, of virtually every dollar earned. 

Of that, significant amounts are budgeted for medical programs for both the aged (Medicare) and the disabled (Medicaid). There are Social Security retirement benefits to pay. And then there is the disability program, which is the focus of this discussion. Some contend that the root cause of financial woe for all that is Social Security is not state workers' compensation, as the authors seem to suggest, but a simple mathematical failure. Critics contend that those who designed the program never balanced the books and that financial failure of the programs was always known and predictable, but was too far in the future to scare anyone. But now the future is here, and someone will have to figure out how to cover the commitments our politicians made. 

The authors seem to discount this contention without further discussion. They seem instead to see state law changes as driving people to SSDI. If this is true, then there could be two logical explanations. First, state laws historically were too lax, providing workers' compensation benefits for injuries and maladies that were minimally, but not significantly, related to work. Thus, primarily personal conditions and their costs were historically "shifted" from personal responsibility or social programs to workers' compensation. Or, it is appropriate for employers to be responsible for the full cost of medical and disability benefits when work plays any significant role in causation, and failure to adhere to this "contribution" standard "shifts" responsibility from workers' compensation to social programs. 

This is a debate, essentially, about the definitions. What is "work-related?" Is it appropriate to use a standard that places full responsibility on a business when most of the cause for treatment and disability is something other than the workplace? When the majority of the cause is not work, should the responsible party be the employer? Or, when the employer is responsible for a significant portion of the cause, is it appropriate to shift the responsibility for all of the care and disability to Social Security? The devil is in the definitions. I suspect that those who pine for federal standards for workers' compensation see some panacea there. Can't we just pay everyone everything they desire? A great solution until the time comes to pay for it. With all of the fine examples of successful socialistic nations in the world, perhaps the question is axiomatic?

It is true that federal standards in this regard, definitions, might provide consistency between all state workers' compensation systems. But, would that consistency put responsibility on the employer or the social safety net? In other words, would those clamoring for national standards be satisfied if the federal government adopted Florida's "major contributing cause?" When I posed that recently to a group, I was assured that I did not understand the issue being debated and was "being silly." The debate, it seems, may not be about consistency but about federal imposition of socialistic standards that place burden on employers. But in that I may be mistaken.

If the responsibility is placed on business, then the cost is not reflected in black and white on people's paycheck. If the responsibility is on the employer, on workers' compensation, then the cost is borne as an expense (workers' compensation insurance premium), and the business sets its prices for goods or services in a manner that covers this expense just like its other expenses (rent, maintenance, supplies, payroll, etc.). And we all pay that cost with each and every purchase we make because it is included in the price we pay. 

If the responsibility is instead placed on the social safety net, then the price (taxes) must be collected or the program will become bankrupt (sorry, is bankrupt). The price must be collected from the paychecks of those that it protects (as Burton and Guo note "regardless of the source of disability"). The price must be collected and that 7.5%/15% goes to 10%/20% or more. And while we can find many who think taxes should increase, it is hard to find people who think their taxes should increase. 

There is too little discussion of the role of the federal government in the financial state of SSDI. While the authors are focused on the definitions in state law, upon which they contend responsibility is shifted from workers' compensation, there is almost no discussion of the changes that SSDI has made in defining eligibility for its benefits. There have been a myriad of changes in that process over the years. In fact, the federal government's own website describes how eligibility has been expanded. Could it be that some of the expansion in SSDI population has resulted from it becoming easier to qualify with an increased set of disability parameters (definitions) in that program?

While the discussion of workers' compensation benefit fluctuation is interesting, it is unfocused. Burton says that when workers' compensation pays less, more people pursue SSDI, and that the inverse is also true: when workers' compensation pays more, or more readily, then less sought SSDI. But this is nothing but mathematics. This explains nothing but a simple truism, when people do not have enough they will seek more. That does not explain, or even begin to explain, the real question of whether workers' compensation or SSDI should be responsible, what the definitions should be. It just says that restriction and constriction shift focus. Interesting, but not really helpful in the real debate. 

Perhaps the most troubling aspect of this analysis is the reliance on self-reported conclusory qualification. The article notes that "one national study, for example, showed that among those in the study group who reported that their health condition was caused by their work," demonstrated certain outcomes in terms of actual qualification for either workers' compensation or SSDI benefits. Even if the definition of causation is consistent (all states used "major contributing cause," or all used some other standard), the determination of whether an injury in a particular case meets that standard or not is a decision for medical experts. 

No validity can be ascribed to analysis that depends on people self-determining whether their injury is work-related or not. This is as baseless as my allowing that audience above to decide if they were or were not computer experts. As an aside, the 25 people I polled, who almost all said they were "experts" were a sixth grade class in which I lectured this spring. These "experts," were allowed to choose whether they would or would not be so defined. And they almost all chose that designation, without any definition of what "expert" means. There was no standard or definition in the inquiry "who is a computer expert." Likewise, we cannot judge the proverbial "cost-shifting" by asking "how many of you with disabilities think they are work-related?"

First comes definition. First we decide what to count. Then, we do not ask subjective and conclusory questions like "how many of you are frequent beach-goers," we ask specific questions using the definition such as "how many of you have been to the beach at least ___ times in the last two years." Analysis that is not fact based, which quantifies personal conclusions and opinions without any empirical measure will almost certainly produce inconsistent and therefore questionable results. 

At the end of the analysis, if everyone agrees that Someone has to Pay, and that this means either Social Security or workers' compensation (that the worker themselves will not pay), then the effect may ultimately be the same. Either through transparent costs, documented on each paycheck, or through hidden costs of workers' compensation premiums, the money is collected, held, and paid to those who we conclude are entitled. The ultimate payer remains the same; it is you and me, the consumers of services. From workers' paychecks or from insurance premiums, the cost will be insinuated in all that we consume. We will pay. 

And so, the devil is in the definitions. What will our standard be for entitlement? How will either program decide who is paid and who is not. Will we admit that both workers' compensation and SSDI have made definitional changes, or will we decry one and ignore the other? Will we compensate disability or impairment? Will we compensate when one cannot return to the work they wish or have historically held, or when one cannot return to work at all?

The bottom line to all of this is simple: THE ANSWERS ARE NOT SIMPLE. There are major policy choices. Having decided long ago (states are celebrating 100 years of workers' compensation; Social Security is about 80; enacted in 1935, benefits were paid starting in 1940) that society would somehow pay the cost of retirement and disability, we now face the hard choices of how much, on what standards, and how? As an aside, perhaps one could argue that a more liberal, inclusive definition for "work related" was logical when workers' compensation was the only game in town at the dawn of the Twentieth Century, but that with the later advent of SSDI, a more restrictive definition is logical? 

One method of analysis might be to point fingers at the definitions used in the various programs, including those in SSDI. Another might be an open debate of the real issues of to whom, in what circumstances, when, and how much. And if consistency is the real goal, then there is room for honest debate about what those consistent standards might be. The right one might be "major contributing cause?" Assuming that consistency necessarily means more or less responsibility for one system or another is presumptive and misplaced. 

No comments:

Post a Comment