UK CLOSED ENDED EQUITY FUNDS

To explore the performance of closed ended equity funds in the United Kingdom during the last 5 years and the discount factor puzzle

To explore the performance of closed ended equity funds in the United Kingdom during the last 5 years and the discount factor puzzle

Data Sources

While the daily market prices for closed end funds are available on many data sets, a historical time series of daily NAVs is more difficult to obtain. Most funds calculate and release their NAVs daily but most standard data sets do not collect the numbers. After some extensive searching, we found one data source, Yahoo Finance that maintains daily NAVs on a large number of closed end funds. The Yahoo NAV data series was started rather recently and many of the funds contained therein do not report pre-2000 data. Accordingly, our data set is somewhat limited.

Index construction and description of variables

Variables

In our empirical analysis, we consider both fund-level and systematic liquidity/ sentiment variables that our model (or the sentiment-based alternative) predicts ought to affect the premium.

Expense ratio and payout ratio

Either increasing the payout ratio, payout, or lowering the expense ratio, exprat, increases the share of the fund's cash flows that go to the investor rather than the manager. As a result, we expect the premium to be negatively related to the expense ratio, and positively related to the payout ratio in the liquidity model. While the types of sentiment model proposed by DeLong et al. (1990) and Lee, Shleifer, and Thaler (1991) do not make explicit predictions about the relation between fund-specific variables and the fund premium, interpreting ?t in our model as a sentiment variable clearly leads to the same predictions as would be obtained in the liquidity-based interpretation. Data on payout are obtained from CRSP, and data on exprat are obtained from S&P Capital IQ.

CEF liquidity

In our liquidity model, the premium is driven by the relative liquidity of the CEF versus its underlying assets. The higher the CEF's trading costs, the lower its liquidity advantage relative to the underlying assets. Everything else being equal, we expect the premium to be negatively related to CEF liquidity, for which we use two different measures, the Roll (1984) trading cost measure, cmdm, and an estimate of the Pástor and Stambaugh (2003) reversal measure of liquidity, gamma. Both measures are obtained from Joel Hasbrouck.

Fund leverage

While our model does not incorporate leverage effects, Table 2 indicates that CEFs make intensive use of leverage. It is therefore important in the empirical investigation to account for any possible effects leverage might have on the premium.25 Data on each fund's leverage, lev, are obtained by interpolating quarterly data from S&P Capital IQ.

Liquidity measures

Both the Pástor and Stambaugh (2003) measure of aggregate liquidity, liq_level2, and the Sadka (2006) measure of aggregate illiquidity, variable_component, are priced liquidity factors. Consequently, one might expect the systematic liquidity of domestic equity, foreign equity, and the types of assets managed by “other” CEFs, to be correlated with these ...