To the Editor:
In the article “A Time for Time Deposits” (by Gilkeson, Porter and Smith, December 2010 AAII Journal), the authors state that investors should use their marginal tax rate instead of their average tax rate. Why not use the effective tax rate instead? Using the effective tax rate would allow investors to adjust for their unique circumstances—presuming minimal changes between years. If there is a big change, this obviously would not apply, and investors would use the marginal rate unless they could effectively determine their new effective tax rate.
James Gilkeson Responds:
When making a single investment decision—in this case, deciding whether to invest in a bank CD or a same-maturity U.S. Treasury security—the proper question to ask is what the income (net, after all costs and benefits, including taxes) from each alternative will be. For taxable income (the CD interest), this net or marginal income will be measured using the marginal tax rate paid by the investor, not the average tax rate, because additional dollars of income are taxed at the marginal rate, not the average rate.
If I understand the idea of “effective tax rate” properly, it takes into account tax benefits (exclusions/deductions) that might be lost/gained as income changes. That’s good. All costs and benefits should be considered. But they should still be considered at the margin, not the average. Note that ideally one would take all costs and benefits into account. For example, holding a CD at an institution might qualify an investor for free services such as checks and ATM privileges.
To the Editor:
I greatly enjoyed the article “Is Now the Time to Add Commodities?” by Conover, Jensen, Johnson and Mercer [December 2010 AAII Journal] and hope to see more of this kind of statistical comparison of investing styles in the Journal. The article shows that a 15% commodity exposure is beneficial during periods when Federal Reserve rates are increasing (“restrictive”), regardless of investment style (value, growth, small, large or momentum.) One thing missing in the article was a precise definition of those five styles.
The Authors Respond:
The return series used in the study are free and available for download on Professor Kenneth French’s website: http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html.
Detailed explanations of the portfolio construction methodologies are also available, and the interested reader is encouraged to visit the site for more details. In general, the style portfolios are formed from the universe of all NYSE, Amex and NASDAQ stocks for which sufficient financial and price data were available at the time the portfolios were formed, and the portfolios were reformed annually (with the exception of the momentum portfolio, which was reformed monthly). The small-cap portfolio includes all stocks that have market caps smaller than the 20th percentile of all NYSE stocks when ranked by market cap. The large-cap portfolio includes all stocks that have market caps larger than the 80th percentile of all NYSE stocks when ranked by market cap. The value portfolio includes all stocks that have ratios of book value of equity to market value of equity (book-to-market) larger than the 80th percentile of all NYSE stocks when ranked by book-to-market ratio. The growth portfolio includes all stocks that have ratios of book-to-market smaller than the 20th percentile of all NYSE stocks when ranked by book-to-market ratio. Finally, the momentum portfolio includes all stocks with a return momentum higher than the 90th percentile of all NYSE stocks when ranked by return momentum.
To the Editor:
Bernard Horn suggests in his December 2010 article [“International Diversification: Why It Still Makes Sense”] that international diversification is still an important portfolio consideration, since although correlations between domestic and international equities rise during global equity swoons, correlations subsequently decrease during the longer bullish intervals.
Closer examination of his Table 1, however, reveals that although correlations do, in fact, decrease during bullish equity markets, the trend of those correlations during the bullish periods from 1970 to 2009 rises from 0.35 in 1970 to 0.47 in 1974 to 0.54 in 1990 to 0.65 in 2002 and finally to 0.85 during the current expansion. It looks to me as if the domestic/international equity correlation has risen quite meaningfully during the past 40 years, even during bullish periods. When one considers the volatile effects of currency fluctuations on equity returns, I derive much less reassurance from a nominal correlation of 0.85, or even 0.65. Is Mr. Horn perhaps mining the wrong data?
Richard W. Lieberman, MD
I am somewhat puzzled by the article on international diversification. The results show that it does pay to diversify. However, the total correlation between the EAFE index and S&P 500 is 0.59. This is a quite high correlation coefficient, which I bet is very significant in a statistical sense (probably p<0.05 or maybe p<0.01). I am surprised at the contradiction between the performance and the high correlation.
Bernard Horn Responds:
From a purely statistical standpoint, Mr. Hausner’s question may be relevant. While the article does not include tests of statistical significance, the correlation between international and U.S. domestic stocks is persistently lower than 1.0, and this is the practical result we are looking for in constructing equity-only portfolios. Further research is needed to determine whether the correlation results we find represent a statistically significant reduction in correlation.
As we know, the variance (or risk) of a portfolio comes down if assets with less than perfect correlation (<1.0) with the portfolio are added to an existing portfolio that is composed of a homogeneous type (e.g., all U.S. stocks). Table 1 shows the calculation of portfolio variance using a simple portfolio consisting of two assets, A and B, where the two assets have different correlations. While the formula is a bit complex to explain here, the results show that as you mix two assets with lower correlation, the overall risk is lower.
Therefore, if international stocks, which in the past had an average correlation of 0.59 with U.S. stocks, are added to a portfolio of U.S.–only stocks, then the variance of the resulting portfolio should come down. Again, in this context we are not making any claims as to the statistical significance of the correlation itself. But if we can generate the same alpha with a lower variancewe can achieve a primary goal of portfolio management. From a pure variance reduction purpose, adding international stocks to your stock portfolio may make sense. Importantly, the persistence of low correlations (i.e., reverting back to the mean) over long time periods also alerts us to the important diversification benefits of international equities. If we then research those individual stocks that have very low correlations with an existing portfolio rather than an entire index, the results can be far more meaningful and are likely to be more statistically significant.
To the Editor:
Perhaps Charles Rotblut could clarify something for me from his article in the AAII Journal December 2010 issue (“How I Find Lower Risk/Higher Reward Stocks”). On page 30, when valuing a company’s current ratio, he gives three points for having a stronger ratio [above 2.0] but only one point when the ratio is above 1.0 but below 2.0. Why is that? I would think having more than double current assets to liabilities would be better, not worse.
To the Editor:
Regarding Charles Rotblut’s risk/reward criteria, are brokerage analysts persons who use the “wild guess” method? I would suggest dropping the earnings estimates criteria and replacing it with one-point scores for revenues and earnings per share in his screen.
Charles Rotblut Responds:
The reason I assign a higher worse score to current ratios above 2.0 is that if a company has too much cash on its books, management may be tempted to make business decisions that are not in the best interest of shareholders. It’s not that high levels of cash are a bad thing, but rather that they can increase the risks associated with the business. In other words, management may feel like it has money burning a hole in its pocket. Separate of that, there is the drag on earnings from cash earning low rates of interest. I would also suggest looking at inventory levels to make sure they are not inflating the current ratio.
I agree with Mr. Spelbring that brokerage analysts are often wrong in their earnings estimates; we see evidence of this each quarter when so many companies exceed, or miss, estimates. However, analysts are usually right about whether earning forecasts should be raised or cut. Often, when several of the covering analysts raise their estimates, business conditions for the company are better than thought. Conversely, when forecasts are cut, business conditions are often worse than thought. This is why I factor in the change in forecasts, as opposed to the actual forecast.