Worries about “political polarization” and of our “post-factual era” impeding political debate in our society have become commonplace. Liberals, in particular, are often astonished at the seeming indifference of their opponents toward facts and toward the likely consequences of political decisions. “Donald Trump has been helped by a conservative-media environment in which there is no penalty for being wrong all the time,” Josh Barro wrote at Business Insider back in October; likewise, Mother Jones’s Kevin Drum more recently wondered: “Why do Republicans tell such obvious lies?”

A common, though apparently ineffective, response to this frustration is to double down by discussing more facts.

When arguing about politics, it is often helpful to construct the best possible version of your opponent’s reasoning — a task admittedly not always made easy by the current administration. In this spirit, though, maybe the issue is not conservatives’ ignorance of facts, but rather a fundamental difference of values. Taking this point of view seems essential for effective communication across the political divide.

Much normative (or value-based) reasoning by liberals (and mainstream economists) is about the consequences of political actions for the welfare of individuals. Statements about the desirability of policies are based on trading off the consequences for different individuals. If good outcomes result from a policy without many negative consequences, then the policy is a good one. When Sen. Kamala D. Harris (D-Calif.) remarked on the Affordable Care Act this spring, for example, she said, “I feel strongly that when we’re talking about our sick, when we’re talking about our poor … and we’re talking about something that would deny those in need with the relief and the help that they need, that they want and deserve, it does put in place a question about our moral values.” In other words, if a policy will harm the welfare of individuals in need, it’s a bad policy.

Meanwhile, much conservative normative reasoning is about procedures rather than consequences. For example, as long as property rights and free exchange are guaranteed, the outcome is deemed just by definition, regardless of the consequences. People are “deserving” of whatever the market provides them with. For instance, Speaker of the House Paul D. Ryan (R-Wis.) seemed to center the idea of unfairness in his argument against the Affordable Care Act: “The idea of Obamacare is,” he said, “that the people who are healthy pay for the people who are sick.”

As an example of how these value differences might matter more than facts, consider the example of bequest taxes, labeled “estate taxes” by liberals and “death taxes” by conservatives. A liberal might invoke various empirical facts: Bequests are very unequally distributed (more so than income or even wealth), bequests primarily benefit those who are already rich, the intergenerational correlation of economic status is very high in the United States and partly driven by bequests, the responsiveness of savings to changes in the bequest tax rate is low, etc. From this, our empiricist liberal might conclude that bequest taxes are an effective policy instrument, providing public revenue and promoting equality of opportunity.

The conservative addressee of these facts might now just shrug her shoulders and say “no thanks.” Our conservative likely believes that everyone has the right to keep the fruits of her labor, and free contracts of exchange between any two parties are nobody else’s business. She will consider someone who has worked hard their whole life, has been frugal and saved their income rather than indulging in consumption, and has raised children to whom he or she intends to pass on the product of his or her achievements. Such a person has lived the model of a moral life. Taxing bequests thus means punishing moral behavior, the exact opposite of what the government should do. And no empirical facts regarding the distribution of bequests or their responsiveness to policy changes affect this assessment in any way.

Exasperated, the liberal empiricist then bemoans the post-factual state of contemporary political discourse. What else does she possibly need to do beyond a plausible empirical analysis of both the status quo and of the outcomes of relevant policy counterfactuals? How can the conservative have read these facts, yet remain unmoved? These differences might make arguing seem impossible. How can our liberal advocating bequest taxes argue with our conservative?

There is no silver bullet, but there are several options, each with its own risks.

A first option is to accept the conservative value framework, but focus on children instead of parents. Consider a child born to rich parents who has never worked hard but indulged in gratuitous consumption in the expectation of receiving a rich inheritance. Such a person is not “deserving” in terms of the ethics of rewarding work; to not reward such immoral behavior, we need to tax bequests.

A second option is to explicitly argue for a liberal value framework. A bequest tax (used to lower taxes on lower earnings) would benefit the majority of people who will never inherit much, but would cost the small minority of rich people who will. We should evaluate the policy based on its effect on individuals, and assign a higher weight to the majority of less wealthy people.

A third option is to challenge the conservative value framework. In a modern society based on a complex division of labor, nobody can be said to consume only the products of their own labor. We rely on social institutions including markets and governments to provide us with all the goods we consume, and absent a theory of just prices (which present day conservatives don’t have) there is no sense in which we are entitled to specific terms of exchange.

Not all political debates are so clear cut as this example. But they all require both facts and values from both liberals and conservatives. Instead of bemoaning other’s intransigence and ignorance of facts, it would often be more productive to reflect on our different value assumptions, and to devise corresponding ways of arguing.