Many closed end fund managers use leverage with the objective of enhancing returns to shareholders. It is probable that few investors question the power of leverage. Surely leverage increases return when the market winds are at your back. It is, of course, a two-edged sword and in down markets leverage should lead to higher losses.

I have started a series describing how an income investor can build a high-income portfolio sufficiently diversified by asset class to smooth at least some of the volatility inherent in high-yield investing (here). The second article in the series explores the role of leverage in closed-end funds (here). I was a bit surprised to find that in the two categories I examined (high-yield fixed income and domestic equity) leverage has provided no meaningful benefit across the categories. This intrigued me sufficiently that I decided to delay a bit the third article in the series to take a look at how widespread this relationship between leverage and returns might be.

My approach was to download data from cefconnect.com for all closed-end funds except non-taxable muni bond funds. The results are through 30 June 2014. This returned 391 funds; two-thirds (260) of these employ leverage. I plotted distribution rate and one-year price and NAV returns against leverage to determine the extent to which leveraged correlates with these return parameters.

The results are shown in the charts below. I added linear regression trend lines for the entire data set (in black), and for funds with less than (red lines) or greater than (green lines) 20% leverage. One can readily see that increasing leverage adds nothing to performance. To the extent there are any trends at all, they are slightly positive with regard to distributions and slightly negative for the share-price performance metrics. But even those slight trends are largely illusory. By looking at the r^{2} values included on the charts, one can see that leverage does explain about 10% of the variation for distributions. At the higher levels of leverage (>20%), however, it has no correlation at all with distribution rates (r^{2}=0.0003). For share price returns, leverage is negatively correlated but only to the extent that explains less than 3% (i.e. r^{2}<0.03) of the variation in market price or NAV at the level of all funds, or for funds with greater than or less than 20% considered separately.

Extending the analysis to total return (market share price % change plus distribution %) for the year, there is no correlation at all.

All of the above only considered leverage funds. How do they compare with the other third of CEFs, those which do not employ leverage? To find out, I considered funds without leverage and cohorts levered at 5% increments. Those results are shown below and are consistent with the correlation analyses. For share-price performance the lowest cohort of levered funds (<5%) is the top performer. For distribution rate, there is some small benefit to leverage, although the pattern is, as the correlations suggest, inconsistent.

And, for total return (market price):

For total return, the most important metric, for only the lowest (<5%, 13 of the 391 funds) and highest (>40%, 7 of the 391 funds) cohorts do the leveraged funds beat the unleveraged funds, and this by only a trivial amount (23.1% for unleveraged, 24.85% for <5%, and 25.2% for > 40%).

What to make of this? Isn't leverage supposed to enhance return in up markets? Of course, the other side of that coin is that leverage will increase losses in down markets as well. Perhaps the trends are an artifact of lumping funds across the CEF universe. To determine if this is the case, I separated funds by category, using the cefconnect.com categories. I performed the same linear correlation analysis on funds from every category that contained at least 10 levered funds. Here, I used the sum of the year's distributions and change in market price to represent total return. The charts below show the results.

Here again the black lines represent all funds in the category. Red lines show funds with less than 20% effective leverage and green lines show funds with greater than 20%. I've not included the r2 values, but they are, with few exceptions, <0.02.

There are categories that did benefit from leverage over the past year, notably mortgage bond funds, multisector bond funds and global income funds. By contrast, domestic equity funds show an inverse response to increasing leverage which is particularly surprising when one considers the strength of the domestic equity markets for the trailing twelve months. In the case of senior loans, the negative correlation with leverage is consistent with the fact that this category did not have a good year. In the market charts report from Oppenheimer for the quarter ending June 30, senior loans is shown to be the worst performing fixed-income category YTD. Under such conditions, increasing leverage would be expected to drive decreasing returns.

So far everything has been for one year. I understand that this is an extremely limited time frame and difficult to impossible to use for generalizations. There's a reason I didn't go beyond one year. Simply, I do not have reasonable access to longer term data in a form that I can use without spending untold hours of manual entries. The primary database for CEFs, cefconnect.com, does not return distributions for more than one year other than for individual names. As distributions are a major part of CEF total return, it is almost meaningless to leave them out. But this is the best I can do. A second potential problem with this approach is that it assumes current leverage represents the longer term values for the funds; there is no source for past values of leverage in these funds.

With those caveats in mind, here are the results.

There is nothing in the 3-year analyses to suggest that the 1-year results are an anomaly. Similar trends appear. There is essentially no meaningful correlation between effective leverage and share value based on either NAV or market price. Leverage accounts for < 2% of the variation in NAV and <0.7% of the variation in market price. There is a slight trend toward a positive correlation with effective leverage at low values (<20% leverage). While the correlations are still minor, this trend has the highest r^{2} values seen (0.18 and 0.13) suggesting that leverage may make a small, but meaningful, contribution to returns at low values for leverage. By contrast for the higher leveraged funds (>20%) there is a small negative correlation between share value and leverage.

I have not dissected the 3-year results by category, but I will note that the general equity CEFs tend to fall on the lower end of the leverage scale (see 1-year chart above) although the one-year results for this category showed a surprisingly strong negative correlation.

This is the third look I've had at leverage in closed end funds. The first was a survey of 128 muni bond funds (here). I was surprised then by the relatively low correlation between distributions and leverage; I had expected a stronger correlation. But for that group the leverage to distribution correlation returned an r2 of 0.37 which is much higher than the correlations for this set of taxable funds. At that time I asked "Is that adequate return for the leverage risk?" And answered, " Not sure, but it seems low to me." In retrospect, it looks pretty good.

More recently, I had a look at leverage in CEFs and returned to the role of leverage in driving returns (here) for high-yield income and two equity categories. The almost meaningless contribution of leverage to returns in those categories drove me to take this more comprehensive look. The current results would indicate that lack of return for leverage risk is not limited to a few categories, but is more widespread, at least for taxable income and equity CEFs.

Clearly, leverage is no substitute for good management. Some managers are able to produce results at low levels of leverage well in excess of what others generate at high leverage. This indicates to me that leverage risk is an important consideration, perhaps more important than I had realized, in selecting among closed end funds. Before taking on high leverage risk, I will want to assure myself that that risk will come with a commensurate return.

In summary, let me say that as an investor I accept that leverage can cut both ways, but these results seem to indicate that it may not be a two-edged sword at all. The sword may have only a single edge, cutting on the downside, but blunt when moving up.

**Disclosure: **The author has no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it (other than from Seeking Alpha). The author has no business relationship with any company whose stock is mentioned in this article.