July 2004

Why dividend yields matter


Our strategy for investing in Japan today and over the last three years has been to identify and exploit the distortions in share prices caused by cross shareholdings across the market. We do not expect an end to the bear market, at this stage. We still think an elimination of these distortions is a necessary precursor to a sustainable recovery in the economy and ultimately the market and only then will there be a lasting opportunity to profit from a bull market in Japan. As these distortions still exist today we view the current recovery in prices from April 2003, in a similar light to the rallies in the market in 1992-1993, 1995-1996 and 1998-2000.

Our strategy to exploit inefficiencies is based upon an analysis of the changing motivations for owning shares in Japan. In the past the majority of shares were owned by corporations and financial institutions for the prime purpose of fostering business relationships for mutual benefit. More recently these investors have needed to sell down these cross shareholdings, in order to release capital to support their own businesses and in doing so have passed ownership to portfolio investors, like us. Portfolio investors have no other motivation in owning shares, than to achieve financial returns. Financial returns are an amalgam of tangible returns from dividends and capital gains. Following 14 years of falling equity prices in Japan and with it the elimination of most capital gains and the rapid aging of the population, bringing ever closer the need to fund pension liabilities, the clamour for tangible income has never been greater. This is why dividends matter. This change in ownership has been going on for some time but only slowly. The cross shareholding ratio (including life-insurance companies) peaked at 56%1 in March 1986 and, we judge, could ultimately decline to as low as 10%2. In the last three years the transfer of ownership has accelerated with the ratio falling from 41%1 to 33%1 and with it the demands from new portfolio investors to improve returns increase. The effects of this are both positive and negative and are beginning to be more evident in the market as the change progresses.

The positive effect is the pressure Japanese managements feel to improve returns on equity. This not only focuses management attention on the key elements of the income statement, revenues and costs but also on the asset base, which for many companies remains bloated. Cash generative companies have been forced to consider how best to deploy their free cash flows, recognising that paying higher dividends not only provides a greater tangible return for their new shareholders but also helps the problem of achieving stable or higher returns from an ever expanding asset base. Share buybacks have been particularly popular as companies are able to absorb some of the overhang of supply from their exiting relationship shareholders while putting excess capital to work at the same time. We have been keen to buy long-term holdings in such businesses. The majority of them are stable cash generative franchises with free cash flow yields averaging 5.5%3 and dividend yields averaging 1.3%3. We believe a combination of higher dividends and share buybacks will help narrow the difference in these yields and at the same time improve the internal rates of return in the businesses. How much they narrow and at what speed will crucially determine the degree of risk in the shares over the short term. The faster dividends are raised, the less downside in the shares. If our portfolio of shares yielded 4% we think there would be minimal downside to the current prices, in all but the worst of market circumstances. This year alone we note that 40% of the companies we own in our long portfolio have raised dividends materially due to an explicit change in their payout ratios. As a result, our long portfolio has significantly outperformed the market. This is best illustrated by the performance of the Close Finsbury Japan Fund, which we have been managing since the end of January. Its portfolio emulates that of our long strategy in the Lindsell Train Japan Fund. The NAV is up 16%4 versus a market up just 9%4 in 5 months to the end of June.

Unfortunately as much, if not more, of the effect of this change in ownership is likely to be negative rather than positive for the Japanese stock market as there are relatively few companies with the cash generative characteristics that we seek. Most quoted business have a history of periodically resorting to raising debt or new equity capital in order to fund growth opportunities or simply to maintain the existing business. Even worse we can report on 12-year records of many companies that have not generated any cumulative free cash flow for their owners and have had to raise new equity and debt many times over to keep the business going. Most are saddled with unsustainable ratios of debt to equity. How is it, you might wonder, that these companies have any value? Not only do they have value but also their market capitalisations are often far higher than similar (often less leveraged) businesses in other parts of the world, whilst the dividend yields are lower. These companies exist because, historically, their major shareholders, typically banks and the providers of capital equipment, owned the bulk of the shares and viewed the continued existence of such value destroying companies as sources of business. Such relationship investing perpetuated a broadly based misallocation of capital that has been part of the cause of Japan’s historically low capital output ratio. Compared to the USA three times as much investment in Japan produced the same unit of output. This is one reason why it is essential that Japan dispense with these outmoded practices in order that the economy can compete effectively with its international peers. By definition, most of these cash consumptive businesses are cyclical, highly leveraged and capital intensive. To illustrate, three short positions in the Lindsell Train Japan Fund, Aeon, Nidec and Taiheiyo Cement all have 12-year records of investing more cash in the business than they have been able to generate from ongoing operations. In their case the extra financing has come from banks or bond investors. Following cutbacks in their investment budgets, this year and into the near future, we expect the businesses to generate some free cash flow. Aeon the largest supermarket group in Japan trades on a free cash flow yield of 1.5%5, a dividend yield of 0.5%5 and has net debt 1.55 times its equity. Nidec, a growing company specialising in components for computers and other electronic equipment, trades on a free cash flow yield of 1.6%, dividend yield of 0.3%5 and has net debt 50%5 of its equity. Taiheiyo Cement trades on a free cash yield of 2.7%2, dividend yield of 1%5 and has net debt 35 times its equity. For the dividends of these companies to yield 4%, the level at which the downside risk in the shares might be presumed to be minimal (see example above), their share prices would have to fall 63%, 60% and 33% respectively. This makes the heroic assumption that all free cash flow would be distributed to shareholders, which in the case of Aeon and Taiheiyo in particular would be most unlikely as their bankers and bond investors would, in all probability, expect some to be used to pay down debt to reduce leverage. In recognition of this Aeon and Taiheiyo’s dividends have been unchanged since 1997 and are expected to remain so in the foreseeable future. In our view such companies will eventually derate and be valued more like UK low quality ‘yield’ stocks. Our short portfolio in the Lindsell Train Japan Fund is full up with companies with similar characteristics.

In the mid 1970’s when dividend yields were 3%6 individual investors owned 35%6 of the market directly. Since then yields have fallen to a low of 0.4%6 in 1989 and then risen to 0.9%6 today. Now individuals own less than 20% because dividends matter to them. Unlike the USA, Japanese individuals have only 5%6 of their total financial assets in shares directly and only a further 6%6 invested indirectly through mutual funds, insurance and pension plans. US investors have as much as 48%2,7, invested in shares, both directly and indirectly. Double taxation may have restrained dividend payouts in Japan but recent changes to encourage share ownership have cut withholding taxes to 10%. There remains huge potential for individuals to invest a greater percentage of their savings in equities but like portfolio investors they seek tangible financial returns of which dividends form a crucial part. Japan Real Estate Investment Trusts (‘JREIT’) were successfully sold to individuals when they were first launched in 2002 at yields of 5%. Yields have subsequently fallen to 4%. We think dividend yields of shares need to rise to this order of magnitude to attract an individual following, which means payouts have to rise, where they can, or prices have to fall.

Michael Lindsell
July 2004

1. Tokyo Stock Exchange, Daiwa Securities
2. Lindsell Train Estimates
3. Lindsell Train
4. Close Finsbury Asset Management
5. Bloomberg
6. Tokyo Stock Exchange
7. KBC Securities