It is conventional wisdom to hold an integer multiple of monthly salary (at least three to six months and perhaps as much as twelve) as cash or cash equivalents in an emergency fund (EF). The purpose of an EF is to allow one to weather an interruption in cash flow (e.g. due to loss of employment) or other financial emergency (e.g. major unexpected expense) without tapping one’s retirement or other dedicated or event-specific investments.
Nevertheless, the financial literature suggests only a minority of American households meet even the weakest test of EF adequacy. This begs the question: which set of Americans are behaving rationally, those with or those without EFs?
I’ve asked similar questions on financial forums and until recently had not received an adequate explanation. If EFs are rational one ought to be able to justify them on the grounds that the borrowing costs they avoid are larger than the opportunity cost of holding low interest EF assets (cash and equivalents). I wanted to see the study that showed this or the opposite.
Only recently did someone point me in the right direction. Should Households Establish Emergency Funds? by Charles Hatcher (Financial Counseling and Planning, 2000) addresses this very question in two ways.
The first is a comparison of the opportunity cost per year of having an emergency fund (characterized by the difference in rates of return between a more aggressive investment and that of the EF cash equivalent holding, denoted by the liquidity premium) with its per-year benefits if an emergency occurred (characterized as the avoided borrowing costs). The result is a simple expression for the minimum probability of an emergency in a given year such that holding an EF is rational. (Hatcher assumes that the EF is equal in value to the cash needed in an emergency so the results are independent of EF size.)
Even with borrowing costs as high as 18% APR (typical of some credit cards but much higher than a home equity loan) and a liquidity premium as low as 2% (well below the expected spread between equity and cash holdings) the probability of an emergency must exceed 21% to justify an EF. That’s about one emergency every five years. Of course for lower borrowing costs or higher liquidity premiums an EF is only rational if an emergency is expected with even greater frequency.
Next Hatcher simulates 40 year life-cycles of two individuals, one with an EF and one without. He varies many of the relevant parameters (borrowing costs, liquidity premium, rate of emergency) and asks, under what conditions does one or the other have higher net worth at the end of the simulated life? Again, it turns out that the EF holder is better off only when borrowing costs are very high relative to liquidity premium and emergencies are fairly frequent (e.g. one every four years).
Hatcher concludes with the right first question about EFs: not how big should they be but who should have one? His results suggest that only those with very high borrowing costs and/or with expected high probability of an emergency should consider an EF. For the rest of us, an EF is not rational. It is not financially optimal.
It is worth noting that Hatcher is not entirely alone in his view of EFs. Bi and Montalto also suggest that a reasonable alternative to EFs are home equity lines of credit (Emergency Funds and Alternative Forms of Saving, Financial Services Review 13:93-109, 2004). (Counter argument: equity lines have to be established in advance of an emergency such as job loss and can be revoked.) Bhargava and Lown suggest diversifying funds that would otherwise constitute a cash (or equivalent) EF holding into more risky and higher return asset classes (Preparedness for Financial Emergencies: Evidence from the Survey of Consumer Finances, Financial Counseling and Planning 17(2), 2006).
So, my question has finally been answered. EFs may not be as universally rational as is suggested by some. I have to admit, I never thought so. My EF has always been on the lower end of what is normally recommended for this reason (also, I have pretty good job security). Nevertheless, I do have an EF, and I’m not giving it up. In part it serves as a cash flow buffer, which I need anyway given variations in monthly expenses. The rest just helps me sleep at night, a factor not included in Hatcher’s analysis.
