Watch out for funds carrying around a large cash cushion, as this is often a tax liability.
Despite the complexity of closed-end funds, CEF investors are generally a pretty savvy bunch. Surfing around on the Morningstar Discussion board, it's evident that identifying red flags, such as return of capital, and checking distribution stability are almost second nature. Nevertheless, misconceptions continue to persist, with one of the most prevalent being the oft-cited data point Undistributed Net Investment Income (UNII).
As the mouthful of a name suggests, UNII refers to the amount of investment income earned by a fund (net of fees) that has not been distributed to shareholders. For example, if a fund earned $1 million in net investment income (NII) for a quarter, but only distributed $900,000 to shareholders, its UNII balance would increase by $100,000. Should the fund earn $850,000 in NII the next quarter, it could dip into its $100,000 UNII balance instead of cutting its distribution. Accordingly, many investors regard UNII as a "cushion" for distributions, where a large balance gives managers extra breathing room to smooth out payments. It's easy to fall into this line of thinking. After all, what's safer than a cushion? This may be true for investors concerned about a steady stream of income, but it misses the broader effect on total returns.
To use a more apt metaphor, UNII is like body fat. From a dietary perspective, body fat serves the important function of storing calories. For hunter-gatherers uncertain of their next meal, a little extra body fat can tide them over long enough to find some food. Today's calorie-rich society distorts this physiological mechanism, however, as the norm has switched to three steady meals per day, many of which consist of several hundred calories. As a result, overnourishment is now a greater health problem than undernourishment in many developed countries. Some estimates peg a third of Americans as clinically obese, with heart disease the number-one cause of death.
The relationship between UNII balance and body fat may seem tenuous, but there is a clear parallel: Small amounts are healthy for the proper functioning of an organism, but larger quantities can be detrimental.
Is UNII Useful?
Let's walk through an example to help illustrate the point. Consider a fixed-income fund with $100 million in net assets and 1 million common shares outstanding (net asset value, or NAV, is $100). As in the example stated above, the fund earns $1 million from coupon payments in the first quarter. With a quarterly distribution of $0.90 per share, the fund pays out $900,000, and the remainder goes into the UNII balance (which is now $100,000, or $0.10 per share).
The important point here is that the fund's NAV increases to $100.10 per share, and the extra $0.10 is reinvested in income-producing assets. 
From an operational standpoint, this extra $0.10 does, in fact, provide a bit of breathing room. If the fund only earns $850,000 the next quarter ($0.85 per share), it can take the extra $0.05 out of its UNII instead of cutting the distribution. From the shareholder's standpoint, doing this is essentially the same as returning capital: Investors would continue to reap their steady $0.90 distribution payment, but the NAV would decrease to $100.05 per share as a result. Instead of realizing a $0.05 drop in income, they realize a $0.05 drop in NAV.
Remember the Tax Implications
Taking tax consequences into account, returning UNII actually can be worse than returning capital. Let's say that an investor buys the fund in our example at NAV ($100.10 per share). If the fund returns $0.05 of capital in its next distribution, NAV goes down to $100.05 and the investor receives a $0.90 distribution instead of an $0.85 distribution. Aside from any confusion that may arise from doing this, the investor would be no worse off receiving a $0.85 distribution because the NAV would stay at $100.10. Alternatively, let's say the fund returns $0.05 of UNII. As in the previous scenario, NAV decreases to $100.05 and the investor receives a $0.90 distribution instead of an $0.85 distribution. The big difference, however, is that this extra $0.05 is taxed as income, whereas return of capital is not taxed.
In other words, investors that buy high-UNII funds are merely taking on a tax liability.
As with return of capital, the premium or discount determines whether the return of UNII will be detrimental to shareholders. If an investor in the 35% tax bracket buys a fund with a $0.40 per share UNII balance, they should keep in mind that buying the fund entails a $0.14 tax liability. If the NAV is $10.00, this means that the fund is only attractive when trading at a 1.4% or greater discount, all else being equal.
One data point that has become popular among many users on our discussion board is the "UNII trend" (i.e. whether the UNII is getting larger or smaller). This is a functionally useless metric, and muddles the overall point. The important thing to pay attention to is whether a fund is underearning, overearning, or just earning its distribution, and whether this is likely to continue. UNII growth or shrinkage is just a consequence of this.
The takeaway is that UNII should be treated in a similar manner to unrealized capital gains and avoided when possible, rather than sought after. Because municipal-bond income is not taxed, muni funds are the exception. In this context, UNII should just be ignored entirely.
 In real life, a fund probably would keep a portion of this in cash for the sake of distribution payments.