发帖
雷达卡
[30 marks]
The following question is based on thearticle by Kamakshya Trivedi and Garry Young (henceforth TY) “CorporateCross-Holdings of Equity, Leverage and Pensions: Simulation and Empirical Evidencefrom the UK” Economic Journal, 116, pp.C190-C208.
The purpose of the paper is to seewhether the behaviour of a company’s share price depends upon the size of thecompany’s pension fund. The relevant behaviour of each company’s share price ismodelled by the company’s “beta” (a measure of the co-movement of the company’sstock return with the market return).
TY hypothesise that a company’s beta will be larger if the company has alarger pension fund (relative to the size of the underlying company).
They conduct the following regressions:
wherePenAssets is the measure of company i’stotal assets invested in equity in the pension fund, expressed relative to thesize of the company (measured by stock market capitalisation).
PenLiabilities is the sum of debt andreported liabilities of company i’spension fund also expressed as a percentage of the size of the company.
The additional explanatory variables(the Market-to-Book Ratio and the Log of Total Assets) are included by TYbecause previous studies have found these to be important determinants of stockmarket betas.
A summary of TY’sresults are reported in the table below.
RHS variable | (1) OLS | (2) OLS | (3) RWLS |
| Pension equity assets as % of market capitalisation | 0.2234 (0.1054) | 0.2585 (0.1020) | 0.2452 (0.1105) |
| Pension liabilites + debt as a % of market capitalisation | 0.0534 (0.0159) | 0.0364 (0.0172) | 0.0403 (0.0246) |
| Market-to-Book Ratio |
| –0.4642 (2.0723) | 0.9109 (2.9291) |
| Log of Total Assets |
| 5.0546 (1.9835) | 5.1043 (2.0583) |
| R-squared | 0.129 | 0.156 | n.a. |
| Number of Observations | 216 | 216 | 216 |
In all threeregressions the dependent variable is the company’s “beta”. Heteroskedasticitycorrected standard errors are reported in parentheses.
TY’spaper does not report summary statistics for the variables of interest.
However, in an earlier discussion paperthey reported data for a subset of 94 companies, from which it is possible toextract information on pensions equity assets.
Some data on company “betas” from a follow-up paper byRachel Ma are also reported in the following table.
Variable | Mean | St dev |
| Pension equity assets as % of market capitalisation | 27.9 | 55.4 |
| Estimates of company betas (FTSE 100) | 0.966 | 0.464 |
(a)
[15 marks] Referring to TY’s OLSresults in columns (1) and (2), do the data support TY’s hypothesis that acompany’s share price behaviour (ie its “beta”) depends partly upon its pensionfund?
Discuss both the statistical and economic significance of the results.
(b)
[6 marks] Do TY’s results supportprevious analysis that the Market-to-Book Ratio and Log of Total Assets areimportant influences on companies’ betas?
(c)
[5 marks] What can we learn from theregressions (1) and (2) in the table about the correlation between the pensionvariables and the market-to-book and size variables?
(d)
[4 marks] Column (3) of the tablereports the results of an alternative estimation technique (RWLS, or“ReWeighted Least Squares”) which attempts to remove the effects of outliers onthe parameter estimates.
Do theresults in the Table and suggest that outliers were much of a problem?
cheche彻底没有啥概念。希望大牛给小弟指点，指点。。
dche
京公网安备 11010802022788号
