The Separation of Intelligence and Control: Retirement Savings and the Limits of Soft Paternalism

Abstract

“Soft paternalism” is in vogue among academics and lawmakers, but too much is being asked of it. This Article studies soft paternalist techniques — including nudging and disclosure — that have been used in the employer-sponsored retirement system. Defined-contribution retirement plans represent an ideal test case for libertarian paternalism: there has been extensive experimentation, and it has often been held up as a success by nudge advocates. In particular, this Article focuses on investment allocation decisions in retirement portfolios, and suggests we should be skeptical of the ability of soft paternalism to improve those decisions. When a domain is rife with conflicts of interest — as in the allocation context — soft-touch strategies fare poorly. Since our tax-incentivized retirement system has paternalistic roots, we should more readily consider direct regulation of investment options available to retirement accounts. The migration of American retirement savings from centralized, risk-pooling structures (Social Security and pensions) towards individual retirement plans (401(k) plans and other tax-favored, individually managed accounts) had collateral consequences. In particular, the responsibility for making complicated financial choices was redistributed to the individual saver — who typically lacks the knowledge and sophistication to make such choices. The result has been that many savers make costly mistakes in investing their portfolios. In response, academics and policymakers, most formally through the Pension Protection Act of 2006, have turned to a variety of typical “soft” remedies, including nudges designed to improve investment decisions by allowing employers to automatically direct employee savings into certain default mutual funds.This Article argues that nudges have failed and will continue to fail in improving the allocation of retirement portfolios, because of problems that are common in many nudge programs. First, nudges rarely consider the ability of third parties to counter-nudge or to weaken nudge outcomes. Conflicts of interest are pervasive in the mutual fund and retirement industry, and those who accept the nudges are being pushed into a category of funds of dubious merit, and which appear to be worsening as institutions seek to exploit the default. Second, nudges are often loosely connected, or not connected at all, to the cognitive problems they seek to remedy. In the retirement allocation context, the nudge acts as a weak mandate for a substantive preference, rather than as a corrective for investors’ cognitive biases. Finally, nudging often asserts autonomy — taking an agent’s preferences seriously — as its central goal. But the claim that the retirement allocation nudges respect savers’ preferences is problematic as a descriptive matter, and illogical as a normative matter in a domain that is already a government-sponsored, tax-advantaged, paternalistic means to encourage retirement savings.