A brief discussion of transaction costs, financial econometrics and market forecasts.

1. Literature review example

(This is a section of a larger body of work)

A strong assumption behind the efficient market hypothesis and the CAPM is that financial markets are perfect. In other words, trading in financial securities is costless, untaxed and investors can effortlessly, as well as instantly, obtain and process information. These assumptions are undoubtedly violated in practice. Thus, it is plausible that the size and the value effects vanish after costs associated with investing are taken into account, and thus are not anomalous at all.

Markets can be imperfect in a number of ways. Direct costs of trading are often ignored in asset pricing studies, but they may be extremely important (Stoll and Whaley, 1983; Alexander, 2000). According to Amihud and Mendelson (1986), the ease with which a share can be sold is a source of risk that is not captured by the static CAPM, and a proxy that captures assets’ liquidity should reliably predict returns. Merton (1987) adds to this point. He notes that gathering and interpreting information is costly and since these search costs are not uniformly distributed in the cross-section of firms, a parameter that measures firms’ recognition among investors should predict returns. Interestingly, Hou and Moskowitz (2005) show that the information cost hypothesis of Merton (1987) is different from Amihud and Mendelson’s (1986) illiquidity story, as these two market imperfections affect returns independently of one another. Consequently, in this section it is shown how recognition of transaction costs, information costs and illiquidity risk in asset pricing augments the understanding of the size premium. A brief discussion of the impact of market microstructure effects on the value premium is left to the end.

2. Financial econometrics example

(This is a section of a larger body of work)

Section X outlined the importance of a comprehensive setup. I noted, following Rigobon (2001), that omission of macroeconomic variables from estimation leads to a significant bias in coefficients, while ignoring heteroskedasticity leads to over-rejection of the null of no contagion. In order to conduct a robust set of tests, a rich structure is required. Following Rigobon and Forbes (2002), however, I do make three assumptions that eliminate the endogeneity effects in the estimation. First, I choose the US market as the leader country where the shocks originate, as in Bekaert, Harvey and Ng (2005). Second, I also assume it to be the world factor. Third, I assume that there is no feedback effect between the idiosyncratic shocks to the value premium back to the market. To recap, I motivate equation (4.1) by, inter alia, Zhang (2006), Bekaert, Harvey and Zhang (2005) and Hou, Karolyi, and Kho (2009). For instance, Hou, Karolyi, and Kho (2009) show that a three-factor model of world markets, value factors and a momentum factor[1] is sufficient to explain global return variation. My setup is analogous to the model of Bekaert, Harvey (1997), Ng (2000), Baele (2005), and Bekaert, Harvey and Ng (2005). They use it look at linkages between equity markets, but I extend it to the international value premium.

4.1 Econometric Setup

The market factor is included in the DGP to capture common shocks and to alleviate the omitted variable problem mentioned in Rigobon (2001). It is motivated, inter alia, by Fama and French (1996) who show that a non-negligible portion of variation in the value premium can be explained with the CAPM market factor in the US, and Fama and French (1998) who show that the world market factor co-varies with the world value factor. Following the evidence of partial integration in the markets (Bekaert and Harvey, 1995) and the value premium (Bekaert, Harvey and Zhang, 2005), I include regional and the factors in the specification.

2. Business analysis example

(This is a section of a larger body of work)

Before discussing the forecasts, I show below the long-term guidance of DPZ. I follow the company by forecasting growth in sales using the growth in the same-store sales and the number of stores as relevant metrics. I also make an adjustment for sales in new shops. However, I show that by following these guidelines, DPZ’s global sales growth of 6-10% p.a. looks ambitious.

In summary, I see two segments in DPZ: slower growth in the domestic segment, and a faster growing international segment.

In the domestic segment, given the different economics of the shops (franchised vs. owned) in the home market, and the split between fees and supply chain in the international sales, a degree of granularity is needed. Figure 1 in the Appendix gives the details.

Domestically, I predict a small increase of about 20 shops per annum, all of which are franchised. I assume that the DPZ has stopped aggressive rationalization of company-owned stores. I also assume that there is limited scope for opening new stores due to the market’s lack of growth and market share growth limits, as discussed above.

I model international store growth by predicating a 15% annual increase in new stores opening up to 2017E, resulting in an 11–12% p.a. growth in international stores. By 2017E, this progression takes DPZ to about 80% of its target to double its international presence (of an unspecified time-frame). I believe this can be achieved, but the table below shows that a lot of the upside is in India and Turkey (c. 850 stores), as well as the UK and France (c. 900 stores).

Figure 1 in the Appendix shows the growth assumptions. For both domestic company-owned stores and franchised stores, I assume the same store growth rate of 3% p.a. in 2014E that gradually moderates to 2% p.a. in 2017E. This is on the lower end of the company guidance and below the 5.5% growth seen in 2013.

There are number of reason that the same-store growth seen in 2013 may have been short-lived. First, given that same-store growth is tied to changes (as opposed to level) in consumer confidence, unless consumer confidence continues to improve, same-store growth is likely to moderate. Second, recent growth has been achieved by taking the market share of larger national players and not smaller local and regional players, whose market share in the home-delivery segment has plateaued at 47%. Given that in the US, the home-delivery market is stagnating, achieving growth via market share gains, especially against larger national players, will be difficult. Third, digitalization of ordering (via the internet or mobile apps) has been a net positive for same-store sales in 2013, as digital sales lead to larger order sizes. However, this impact is likely to fade as it approaches saturation. In 2013, 40% of US sales have been done online, and given that DPZ does not disclose that number in its 2012 report, it is likely that the 2012 base was low. Consequently, impact on future growth rates is likely to be smaller.

…

[1] I do not include a momentum factor, as it may capture contagion. The interrelation between global momentum factor and contagion can be left for future research.