Mistaking choices for bias

A rather interesting example of a common phenomenon. Reporters are often susceptible to the ideological orientations of the left leaning academy and prone to believing some things that are simply not true. The fact that few are quantitatively oriented or have formal grounding in statistics and logic makes the situation even worse. The right sees this as evidence that journalists are a fifth column intentionally advocating for left positions. I think it is much more subtle than that. I don't dispute the empirical fact that most journalists skew Democrat and progressive based on their political donations.

I think the mechanism by which this is revealed is not conscious bias. Instead, I think the bias shows up in two forms. The first is in the matter of what gets reported and what does not. They can't report everything and they tend to report on things that appear important to them (and their editors). What remains unreported are often just as important but not pertinent to their particular political leanings. Corruption, hard and soft, is rife in many major metropolises. For the right, that is one among many indications that big government cannot be trusted and those stories are dramatically underreported. For those of a left persuasion, I think the mental framing is that humans are inherently weak and subject to temptation. It is a dog bites man story and therefore not interesting and indeed is underreported.

The second way that the bias manifests is in the journalist's priors, those assumptions they make in advance. The story is reported in the context of those priors which may be either simply wrong or they may distort the factual data.

The instigating article is from Tax Policy Is Widening the Gender Gap by Victoria Bateman. The argument being made is there in the headline. Another gender grievance. But is it true that tax policy is widening the gender gap?

With British politicians, led by Prime Minister David Cameron, making public their tax affairs, the OECD’s new Taxing Wages report could not have come at a better time. For those tired of Piketty-style class warfare, the report comes as a welcome departure. Move over, class warriors: It’s single people and women who really lose out from our tax and benefit systems.

Economists often look at the tax wedge -- the difference between the total labor costs to an employer and the net take-home pay of the corresponding employee. Income taxes are just one of a number of factors that can drive a wedge between the two figures, reducing incentives to employment. The tax wedge calculated by the OECD captures much more: the sum of income tax, social security contributions (paid by the employer and the employee) and payroll taxes, netted for any cash transfers (or “benefits”), all expressed as a proportion of the labor cost. The higher the tax wedge, the greater is the overall tax on labor.

The article is full of interesting empirical information which is reported but not generally elaborated on. The elaboration comes later in the article.

As the OECD notes, where governments choose to tax family income instead of individual income, “the second earner is effectively taxed at higher marginal tax rates than a single individual would be,” disincentivizing paid employment. Where countries provide tax allowances and tax credits to families (both of which are regularly justified on the grounds of equity), they come with a similar adverse effect, as has been the case in Italy.

Quantitatively speaking, the difference in tax between primary and secondary earners is certainly significant. For a family with two children in which the primary earner is receiving the average wage, the tax wedge on the secondary earner (who is assumed to earn two-thirds of the average wage) was 37.8 percent across the OECD in 2014. This compares with 26.7 percent for the primary earner, a difference of over 10 percentage points.

Where the majority of secondary earners are women, this means that women are being taxed more than men, which raises issues of fairness and potentially affects women’s labor force participation. According to a report published last year by the European Commission, fiscal disincentives facing female secondary earners are a particular problem in Belgium, Germany, Slovenia, Portugal and Luxembourg. Three of that list -- Belgium, Germany and Luxembourg -- have a high proportion of women in part-time as opposed to full-time roles. Joint tax filing (or joint elements) in these countries also means that the tax effectively paid by the second earner is dependent on -- rather than independent of -- the income of the primary earner.

It seems that the tax and benefit system is implicitly working against the achievement of gender equality in the workforce. Rather than blame the private sector for exacerbating inequality, it might be time for the state to look more carefully at how its own labor market interventions are impacting the gender gap.

Because, presumably, of her prior assumptions, Bateman fails to make a distinction. It is possible for two apparently contradictory statements to both be true. Based on her own reported information, it is obvious that "The tax policy is gender neutral" is true as well as that "Women are, on average, more heavily taxed than men." The reason is not that companies are discriminatory or that government is malevolent. The reason that both statements can be true rests with patterns of individual choices.

"The tax policy is gender neutral" is true because the government is not taxing by gender but by type of taxpayer condition. In other words, it is not distinguishing between male and female tax payers but between primary and secondary earners, regardless of sex. It is observable for a variety of economic and policy reasons, that the secondary earner is usually taxed at a higher rate.

It is also true that "Women are more heavily taxed than men." This is because, again for a variety of usually social and familial reasons, people make sets of decisions about the nature of the work they are willing to do and that women more often end up, within their family circumstances, as the secondary earners and therefore more heavily taxed because they are secondary earners (not because they are women.) In families where the male is the secondary earner, he will be taxed at exactly the same rate, given the same circumstances, as the female secondary earner.

The more accurate argument, compared to that made by Bateman, is that "On average, women's choice of work patterns, more often lead them to being taxed at a higher rate than men who make different patterns of choices." The more succinct headline might, "Secondary earners are taxed at a higher rate than primary earners."

It would appear that Bateman subscribes to a prior assumption that women are unfairly treated in the marketplace and therefore she reads the gender neutral tax policy as one more burden that working women bear. In fact, the burden falls equally on men and women and only differs because men and women, as individuals and as family units, make career and work decisions that yield the outcome where women are more often the secondary earner.

Because of her priors, Bateman draws the wrong conclusion from the data.

Because she draws the wrong conclusion, conservative readers conclude that this is one more instance of victim advocacy.

I don't think this is deliberate ideological advocacy. It is simply careless reporting, a result of priors and lack of numeracy.

Which is a pity, because there is a lot of useful and interesting information in the article.