Blue Green Line in Header

The Difference Between Accuracy and Bias - Part 2


In the short time since I last posted an insight, the issue of ‘accuracy vs. bias’ has raised its head yet again. Last time it was about something as broad as a debate on the national unemployment rate. This time it is very specific—namely an attack on Beacon Economics’ study of the potential economic impact of raising the minimum wage in the City of San Jose.

While distinctly different issues in terms of scope and personal interest, in both cases accusations of inaccuracy are incorrectly used to paint a result as biased—even as the accuser is clearly guilty of that exact charge. In other words, these commenters are willing to distort reality in order to pursue a political goal. For myself, and Beacon Economics, it reminds us yet again why we jumped into the economics business in the first place.

To refresh your memory, my last post was about Jack Welch’s complaint about the U.S. unemployment rate – which dropped sharply in August and September—much to Mr. Welch’s chagrin. I pointed out that there was a fundamental flaw in his argument that the nation’s unemployment data was being altered to favor the current president over the opposition. The problem was that Mr. Welch went to great lengths to assert correctly that the unemployment rate was an estimate with a fair degree of ‘noise’ in it—and thereby surmised that the results must be biased. A lack of high accuracy (noise) is a data problem. The issue of bias, however, lies with the researcher. The two should not be confused.

This same issue has reared its head again in a scathing response to our analysis of a City of San Jose minimum wage statute that will be on the ballot in November. The response to our study was authored by Professor Michael Reich from the Institute for Research on Labor and Employment at UC Berkeley, and repeated in an open critique by Santa Clara County Assessor Larry Stone.

First, a bit of background. When Beacon Economics was approached by the California Restaurant Association to conduct an impact study on the potential minimum wage hike in the City of San Jose (Measure D), we hesitated to take the project. This was mainly because we knew it would be an estimate that would have some attackable assumptions behind it. The data needed to measure the impact is based on surveys of the local economy that are quite small in terms of their sample size and have a lot of user entered data which carries with it a potential for error.

This isn’t to say an economic analysis cannot be done—but it is a tricky problem that takes a careful eye and reasonable efforts to try and reconcile data from a number of disparate and often contradictory surveys on the composition of the San Jose economy. Simply coming up with an estimate of the number of workers earning at or below minimum wage is, believe it or not, difficult. It can be done—but it will always leave you open to criticism by those who want to discredit you.

Ultimately, we decided to take this difficult project on, as one of Beacon Economics' goals when the company was formed was to provide an unbiased and as accurate as possible view of the implications of important policy issues facing voters across the state and nation. This, sadly, is something that is lacking in many debates where ‘experts’ are highly biased, and rather than perform honest research, play games with numbers to create conclusions that align with their opinions.

Beacon Economics tries hard to avoid this problem in all our work. When we approach a project, we accept only that there are basic costs and benefits to all policy choices and we have to find the appropriate values on both sides in order to lay out, at the very least, the economic consequences. When drafting our Measure D report, we were as careful as possible, given the challenges, to make a reasonable estimate. We even created a high and low value to give a sense of the results.

In the end, we arrived at what I very strongly believe is a reasonable finding: a hike in the minimum wage would have a negative impact on the City of San Jose, but one that could be characterized overall as mild relative to the massive tech-fueled economic engine we call Silicon Valley. In total, we estimated that the City of San Jose would lose 900 to 3000 jobs—many would likely end up going elsewhere in Santa Clara County. This isn’t to say it is unimportant, since certain sectors—think small mom and pop retail establishments and charity organizations—would feel the impact disproportionally.

Still, this moderate result came under attack by Dr. Reich. And candidly he does make one or two good points about the construction of our analysis. The data we both rely on to do our respective studies have issues, and he clearly understands that. Indeed, I had my staff of researchers look through some of his points—and in the future we may well take his advice on how to estimate the impact of such policy issues in a different way—if for no other reason than to consider a alternative approach to creating an answer.

I won’t list his complaints here. We could easily (and pointlessly) go tit for tat and criticize how he conducted his analysis as much as he did ours. This would ultimately be an academic debate on how to best estimate policy impacts that would undoubtedly bore our audience to tears. And in the end it wouldn’t change our overall result—that the hike in the minimum wage would have a negative impact on the City. In short—it’s a debate about accuracy.

Yet it quickly becomes clear that he, to paraphrase, doth protest too much. Dr. Reich’s motivation in his reply had nothing to do with improving the accuracy of Beacon Economics’ results. Quite the opposite, like Jack Welch, he was trying to use issues of accuracy to imply that Beacon Economics is biased in our work, stating in no uncertain terms that our results were flat out wrong. His report, while full of interesting, if debatable, tidbits of knowledge about census data is really nothing more that a hatchet job.

Why? Because Dr. Reich does not come at these issues with an open mind, but rather with deep-seated biases and ideological beliefs. A brief look at Dr. Reich’s CV shows a long history of involvement in what can be termed 'workers rights' issues, and he runs a center that appears to have a strong liberal agenda. The point of Dr. Reich’s analysis is to try and quash research that is ‘wrong’ only inasmuch as it does not subscribe to his vision of world order.

Where Dr. Reich tips his hand and shows his bias is when he fails to acknowledge that there are always trade-offs and, instead, simply asserts that his research shows that minimum wage increases have no impact on employment and economic output. None. And as such there is nothing but benefit to be had by the local economy in hiking the minimum wage – no matter how large or small of an increase.

This kind of nonsense is not worthy of someone who has a PhD in economics. When the government actively works to manipulate any market by fixing prices, there will be negative consequences. We might argue that the benefits are greater than the negative results, either because of social equity issues, or market failures, such as externalities. But to pretend there is no negative impact is simply preposterous.

If you don't believe me – do a simple thought experiment. If there is no negative impact to a higher minimum wage, why not raise it to $12 an hour instead of $10? Or $120 per hour? Or $1,200 per hour? Dr. Reich, when does it become an issue for economic output?

And of course we have to keep in mind that someone has to pay those higher wages—a small mom and pop storeowner, a charitable organization, or a major retail chain hiring a teenager for their first job. And that means someone, somewhere, will spend less even as these workers spend more. At the very least we can argue that when funds are taken from one person and given to another at best the result is a neutral transfer. To claim that a raise in the minimum wage has all positive and no negative impacts is a complete non-sequitor hidden behind a mass of data complaints.

Dr. Riech and his brand of biased policy declarations are exactly why I founded Beacon Economics. I don’t believe that Dr. Reich actually thinks there are no downsides to hikes in the minimum wage. What I believe is that he knows there are negative impacts—but he is comfortable lying about those consequences in order to achieve what he feels is a degree of social justice. In short, the policy ends justify the biased research means, as well as the hatchet jobs he tosses out along the way.

In our view the path to a better world is not through distortions and attack articles—but honest discussions regarding the real trade-offs involved in policy decisions. I stand behind Beacon Economics’ analysis of Measure D, and in all our work we will continue to provide the best possible, unbiased research we can. If we make a mistake, we promise to admit it and look for a better answer. But it will always be an answer that comes from logic, reason, and data – not from a preset political bias designed to con the world into doing something we feel is ‘right’.

Would an increase in the minimum wage have some negative impact on the local economy? Of course it would. Will it be offset by some positive economic impact? Yes, but not enough to create an overall positive economic impact. Is it still worth doing for social equity reasons? That is a questions the citizens of San Jose have to ask themselves when they go to the polls. It is not up to me, or Dr. Reich, to decide for them.

CATEGORY: Economic Policy

Previous article  |  Next article