Tuesday, January 27, 2015

Text Mining

(1) 第一个是pdf,第二个是youtube教程。貌似主要命令是proc hptmine

Enabling High-Performance Text Mining - SAS

https://support.sas.com/.../Tips-High-Performance-Text-Min...

SAS Institute
HPTMINE and HPTMSCORE. This option allows for high performance execution usingSAS® code specification. Both of these high performance text mining  ...



  1. SAS® High-Performance Text Mining - YouTube

    www.youtube.com/watch?v=k5ydJXWf6Ug

    Oct 23, 2013 - Uploaded by SAS Software
    Alan Zhao, in the Analytics Product Management department atSAS, show how to handle very large problems ...
(2) Leone的主页,关于Perl和SAS macro运用



  1. Perl Resources - University of Miami

    https://sbaleone.bus.miami.edu/PERLCOURSE/Perl_Resources.html

    Andy Leone's Perl Page. Below are some programs for, among other things, downloading and extracting data from Edgar using Perl. If you are new to Perl, there ...
有关于Event study算简单的median, mean, std
也有winsorize的macro版本。
各种抓东西的perl程序。

(3) 琳娜给了一个简单的SAS search code

http://www.ats.ucla.edu/stat/sas/faq/index_function.htm

(4) Eric给了一个R的文本

Hands-On Data Science with R Text Mining - One Page R

onepager.togaware.com/TextMiningO.pdf
Nov 5, 2014 - Data Science with R. Hands-On. Text Mining. 1 Loading a Corpus. A corpus is a collection of texts, usually stored electronically, and from which ...

可以总结词汇,成为一个word cloud。用R语言,好神奇!


Dairy 20150127 - Tuesday

(1) 看到John J. McConnell的database,有carve-out,也有spinoffs.

    Krannert School of Management - John J. McConnell


  1. www.krannert.purdue.edu/.../mcconnell...
    Krannert School of Management
    Purdue University ... John J. McConnell - Publications ... Predictability of Long-TermSpinoff Returns · S&P 500 ... Can Takeover Losses Explain Spin-off Gains?
  2. (2) 这公司算是科研机构,先后有130多位full professor给他们工作。spin off了很多高科技公司。随便看看。


  1. Spin-off companies' details | CWI Amsterdam | Research in ...

    www.cwi.nl/general/spinoff-companies
    Centrum Wiskunde & Informatica
    CWI has created a number of spin-off companies: companies founded by CWI employees, with ... Development of database technologies for Business Analytics
(3) 这文很有意思,其中有一个列表,讲了tracking stock的去向,数了数,有8个最后被spin off。值得注意的是,即使被宣告dissolve,不论以什么形式,包括spin off,市场仍旧有正效应,positive CAR。至少说明,在spinoff的CAR中,有一部分效应,不来自于information asymmetry的降低。

Off Track: The Disappearance of Tracking Stocks

T Davidson, J Harper - Journal of Applied Corporate Finance, 2014 - Wiley Online Library
... by Travis Davidson, Ohio University, and Joel Harper, Oklahoma State ... why tracking stocks have
largely stopped being issued, and why so many have been dissolved either through exchanges
of shares, or sales or spin-offs of the business units with tracking stock. ...

(4)这篇文章讲欧洲的spin off在长远期并没有像美国一样,大幅提升公司绩效和市值。

Do spin-offs really create value? The European case

C Veld, YV Veld-Merkoulova - Journal of Banking & Finance, 2004 - Elsevier
... The Krishnaswami and Subramaniam (1999) paper suggests the testable hypothesis that
spin-offs of firms with large information asymmetries are associated with higher abnormal returns
than spin-offs of firms with low information asymmetries. The Habib et al. ...

(5) 这篇文章是model paper,讲了spinoff的别样动机,就是spunoff的子公司,会更可能被收购,所以管理层就会好好努力,所以market value会增加。

A theory of corporate spin-offs

TJ Chemmanur, A Yan - Journal of Financial Economics, 2004 - Elsevier
... Habib et al. (1997) argue that spin-offs improve the quality of the information managers
and uninformed investors can infer from the prices of the firm's traded securities, therefore
leading to an increase in the expected price of the firm's equity. ...

(6) 这篇文章是Ohio State一个phd学生写的dissertation。

[PDF] Disclosure Tone of the Spin-off Prospectus and Insider Trading

W Choi - 2014 - etd.ohiolink.edu
... outliers that perform extremely well. They conclude that “post-spinoff stock returns do ... revealed
to market participants at the time of the spin-off.10 Feldman, Gilson, and Villalonga (2013)
analyze the content of analyst reports written about companies that ...

Friday, January 23, 2015

Dairy 20150123


Dataset List → Compustat (COMPA) → Merged Segments File (Non-Historical) (WRDS_SEGMERGED)


segmerged database 是个什么东东?

两个link,保存一下。

Investigate on diversified firms and their segments using Compustat
http://www.wrds.us/index.php/forum_wrds/viewthread/224/#629


SAS macro to determine number of segments
http://www.wrds.us/index.php/repository/view/34

力力问sample怎么样了?是不是从edgar的S-1,S-3来搜索spin off?

========================================================

SAS file: IPO_SDC20150123

13947 IPOs
1654 not matched in CRSP (力力说,不用去管)

有8个obs,SDC认为是spin off却不是IPO.

ID 28650
CarMax Group(Circuit City Stores Inc)

Carmax (from Circuit City)应该是tracking stock

ID 27455
Digex Inc

  1. Digex, Inc. - Reference For Business

    www.referenceforbusiness.com › Information Technology
    Digex, Inc. began the 1990s as an Internet service provider (ISP). ... Digex Spin-Off: 1999 ... Web Hosting Makes Digex an Acquisition Target: 2000-2001.
In mid-1999, Intermedia spun off Digex as a public company, offering 49 percent of the company's stock to the public.

Read more: http://www.referenceforbusiness.com/history2/40/Digex-Inc.html#ixzz3Pg5GbdfS


SDC数据也不对,SDC写pre-owned % 97.8 to post-owned 95.1%.貌似不符。



SAS macro to determine number of segments

Wednesday, January 21, 2015

Dairy 20150121

(1) 思考neoclassic model 在spinoff中,对于information的作用?

(2) 读读别的merger waves文章

[PDF] The importance of industry links in merger waves

K Ahern, J Harford - Journal of Finance, 2010 - msbfile03.usc.edu
ABSTRACT We represent the economy as a network of industries connected through
customer and supplier trade flows. Using this network topology, we find that stronger product
market connections lead to a greater incidence of cross-industry mergers. Second, ...

我现在对于Neo-classical的理解,基本上就是industry shock。比如regulation, economic, technological shock。这些东西会造成merger wave。反过来,也是造成spin off的原因吧?

=================================================

Decoupling CEO wealth and firm performance: The case of acquiring CEOs

J Harford, K Li - The Journal of Finance, 2007 - Wiley Online Library
ABSTRACT We explore how compensation policies following mergers affect a CEO's
incentives to pursue a merger. We find that even in mergers where bidding shareholders are
worse off, bidding CEOs are better off three quarters of the time. Following a merger, a ...

Valuation waves and merger activity: The empirical evidence

M Rhodes–Kropf, DT Robinson… - Journal of Financial …, 2005 - Elsevier
To test recent theories suggesting that valuation errors affect merger activity, we develop a
decomposition that breaks the market-to-book ratio (M/B) into three components: the firm-
specific pricing deviation from short-run industry pricing; sector-wide, short-run deviations ...

Do envious CEOs cause merger waves?

AM Goel, AV Thakor - Review of Financial Studies, 2010 - Soc Financial Studies
Abstract We develop a theory which shows that merger waves can arise even when the
shocks that precipitated the initial mergers in the wave are idiosyncratic. The analysis
predicts that the earlier acquisitions produce higher bidder returns, involve smaller targets, ...

Merger waves是因为CEOs的嫉妒造成的吗?文章发现:
(1) 早期的Merger小于后来的merger
(2) 在一个merger wave中,后来的bidder return比early bidder return要小(所以后来的CEO因为嫉妒,要获得庞大帝国,一点点蝇头小利,也愣是要merge?)
(3) acquiring firm(应该就是母公司吧)高管的compensation benefit在merger wave的前期,比后期的大。

不管有没有人信这个故事?反正我是不信的。但伦家是The Review of Financial Studies

=================================================

Who makes acquisitions? CEO overconfidence and the market's reaction

U Malmendier, G Tate - Journal of Financial Economics, 2008 - Elsevier
... Abstract. Does CEO overconfidence help to explain merger decisions? Overconfident
CEOs over-estimate their ability to generate returns. As a result, they overpay for
target companies and undertake value-destroying mergers. ...

Corporate Governance and Merger Activity in the US: Making Sense of the 1980s and 1990s

B Holmstrom, SN Kaplan - 2001 - nber.org
... with conglomerates" that had been assembled in the previous merger and acquisition wave in
the 1960s. ... 2. Why Did Corporate Governance and Mergers in the 1990s Look So Different? At
the end of the 1980s, the takeover wave ended. As the previous figures show, takeover ...

(3) 两个会计教授,NYU和Emory写的spinoff paper。有没有underpricing?

The Long‐Run Performance of Sponsored and Conventional Spin‐Offs

A Klein, J Rosenfeld - Financial Management, 2010 - Wiley Online Library
... We also compute the return performances for the conventional spin-offs as well as the parent
firms for both groups of spin-offs. Consistent with recent studies conducting long-run performance
tests (Hertzel et al., 2002), we use the Carhart (1997) four-factor model as our ...

(4) 做表格

Tuesday, January 20, 2015

Dairy 20150120 问题又来了

第一个问题,力力文件中的有些spin off在SDC Equity里,spin off flag并不是Y。深入调查这件事。如果没法解决,应该怎么办?

未解决!

第二个问题,clustering by industry什么意思?Neoclassic又是什么意思?

Neoclassic 的意思是,企业发生并购,重组,是因为economic shock,比如regulations, technology等等。念的Paper是Jarrad Harford (2005, JFE)

What drives merger waves?

第三个问题,Graham有没有写过关于spinoff survery的paper?

未解决!


Monday, January 19, 2015

Classification and Tables

Table 1
Distribution of Spinoff Firms in Time series (NYSE, NASDAQ, etc.)

Table 2
Distribution of Spinoff Firms in different industries

Table 3
Market capitalization of the parents and subsidiaries in the sample
(or pre- and post-descriptive statistics of the financial variables for parents)

Table 4
Descriptive statistics of the financial variables for spinoff firms and control firms

* Total Assets
* Cash flow from operations
* Operating income
* Market-to-book ratio
* Debt ratio
* Unrelated entropy

Total assets is obtained from Compustat and is denominated in billions of dollars. Cash flow from operations is cash flow generated from all operating activities and is measured as a ratio relative to the total assets of the firm. Operating income is sales minus cost of goods sold and other expenses, before depreciation and amortization, and is measured as a ratio relative to the total assets of the firm. Market-to-book ratio is the ratio of (book value of assets- book value of equity + market value of equity) to the book value of assets. Debt ratio is measured as the ratio of short-term plus long-term debt to the total assets of the firm. Unrelated entropy is an index of unrelated diversification of the firm in its operations. For each firm, unrelated entropy is the weighted average of the percentage sales of the various distinct two-digit SIC industry groups within that firm. All variables are measured at year-end of the year prior to the spin-off announcement year.

Table 5
Abnormal returns over selected intervals for parents around announcement dates
-30 to -6
-5 to -1
-1 to 0
0
-1 to 1
+1 to +5
+6 to +30

or maybe for spunoff subsidiaries' IPOs (I never see such a table in any other paper.)

Gross proceeds/(issue size)
Total assets
Offer price
Number of shares
Underpricng

Underpricing is measured as [(MP - OP)/OP] x 100, where MP is the aftermarket closing price of the first trading day recorded by CRSP and OP is the offer price.

Table 6
Panel A
Information asymmetry between parents and control firms

Forecast error
Std. deviation of forecasts
Normalized forecast error
Announcement reaction
Residual std. deviation

Panel B
Information asymmetry for parents before and after spinoffs


The forecast errors are defined as the ratio of the absolute value of the difference between the actual earnings and the forecast earnings to the price per share at the beginning of the month. The before-event forecast errors are measured in the last month of the fiscal year before the announcement of the spin-off. The after-event forecast errors are measured in the last month of the first fiscal year after the completion of the spin-off. The standard deviation of forecasts measures the dispersion in the analysts' earnings forecasts in the month in which the forecast errors are computed. The before-event normalized forecast errors are measured as the ratio of the before event forecast errors to the standard deviation of the detrended quarterly earnings in the five years before the announcement of the spin-off. The after-event normalized forecast errors are measured as the ratio of the after-event forecast errors to the standard deviation of the detrended quarterly earnings in the two years after the completion of the spin-off. The before-event announcement reaction variable is the standard deviation of the three-day cumulative abnormal returns around the announcements of quarterly earnings in the five year period before the announcement of the spin-off. The after-event announcement reaction variable is the standard deviation of the three-day cumulative abnormal returns around the announcements of quarterly earnings in the two year period after the completion of the spin-off. The before-event residual standard deviation is the dispersion in the market-adjusted daily stock returns in the year preceding the spin-off announcement. The after-event residual standard deviation is the dispersion in the market-adjusted daily stock returns in the year after the spin-off completion. The results of the non parametric Wilcoxon Signed Rank test for the difference in the means and the Median Scores test for the difference in the medians are specified in the panels.

Table 7
Panel A
Descriptive Statistics of Management Forecasts (all firms in all the time)
                       Q1              Q2              Q3              Q4
1994
...
2014

Panel B
                           Parents              Sub              Control
Year -1
Year 0
Year +1

Panel C
                                            Parents -1, Parents +1, Sub +1, Control -1, Control +1
Good news forecasts                   n
Bad news forecasts
Neutral forecasts
Total

Good (bad, neutral) news forecasts refer to management forecasts for which the point estimate, or the mid-point of a range forecast, is above (below, similar to) the average of the analyst forecasts issued in the 90 days before the management forecast release date. Management forecasts are classified as neutral news if the absolute value of forecast news (FN, defined as management forecast minus average analyst forecast scaled by pre-release price) is in the bottom quintile of the sample distribution.

Table 8
Panel A
Abnormal returns for parent firms that completed a spin-off in the period 1994-2014, sorted
based on the level of information asymmetry.

                                           Abnormal Return Quartile 1  ...   Quartile 4         Difference
Forecast error
Std. deviation of forecasts
Normalized forecast error
Announcement reaction
Residual std. deviation

Panel B
Abnormal returns for parent firms that completed a spin-off in the period 1994-2014, sorted
based on the level of management forecasts descriptive statistics.
                                           Abnormal Return Quartile 1  ...   Quartile 4         Difference
Management forecasts frequency
Width
Precision
Forecast Error


Width - Forecast width for a range or point forecast, calculated as the difference between the high-end estimate and the low-end estimate, divided by the absolute value of the mid-point of the estimate;

Width is 0 for a point forecast;

Precision - Management forecast precision, defined as forecast width (Width) times 1.

Forecast Error - Management forecast error, calculated as the absolute value of the difference between actual EPS and the management forecast of EPS, divided by the prerelease share price;

Forecast Horizon - Management forecast horizon, calculated as the number of calendar days between the forecast release date and the corresponding earnings announcement date; we use the log transformation in the correlation matrix and regression analysis;


Panel C
Underpricing on the day of subsidiaries' IPOS, sorted based on the level of information asymmetry.

Panel D
Underpricing on the day of subsidiaries' IPOS, sorted based on the level of management forecasts descriptive statistics.

Table 9
Information asymmetry, abnormal returns, and management forecasts descriptive statistics for the sub-samples of same-industry and cross-industry spin-offs.

Panel A: Abnormal returns for cross-industry and same-industry spin-offs
Abnormal returns

Panel B: Information asymmetry in cross-industry and same-industry spin-offs
Forecast error
Std. deviation of forecasts
Normalized forecast error
Announcement reaction
Residual std. deviation

Panel C: Management forecasts descriptive statistics in cross-industry and same-industry spin-offs
Management forecasts frequency
Width
Precision
Forecast Error

Table 10
Information asymmetry, abnormal returns, and management forecasts descriptive statistics for the sub-samples of spin-offs with stated motive to increase transparency.

Panel A: Abnormal returns for sub-samples based on transparency increase motive
Abnormal returns

Panel B: Information asymmetry for sub-samples based on transparency increase motive
Forecast error
Std. deviation of forecasts
Normalized forecast error
Announcement reaction
Residual std. deviation

Panel C: Management forecasts descriptive statistics for sub-samples based on transparency increase motive
Width
Precision
Forecast Error

Table 11
Summary of abnormal returns, information asymmetry, and management forecast descriptive statistics for parent firms that completed a spin-off in the period
1994-2014, sorted based on different motives.

Panel A: Descriptive statistics for sub-samples based on merger motive
                                       Merger Motive            Non-Merger Motive            Difference
Abnormal Returns
Information Asymmetry
Management Forecasts

A merger motive is said to exist if the firm states that the spin-off is intended to facilitate the merger or acquisition of either the parent or the subsidiary with another firm.

Panel B: Descriptive statistics for sub-samples based on regulation motive
                                       Regulation Motive            Non-Regulation Motive            Difference
Abnormal Returns
Information Asymmetry
Management Forecasts

An anti-trust regulation motive is said to exist if either the parent or the subsidiary, but not both, is in an unregulated industry after the spin-off.

Panel C: Descriptive statistics for sub-samples based on tax status motive
                                       Tax status Motive            Non-tax status Motive            Difference
Abnormal Returns
Information Asymmetry
Management Forecasts

Taxable spin-offs are identified from their distribution codes on CRSP.

Panel D: Descriptive statistics for sub-samples based on financing motive
                                       Financing Motive            Non-financing Motive            Difference
Abnormal Returns
Information Asymmetry
Management Forecasts

Financing motive is identified from whether firms state such benefits due to spinoffs.

Table 12 Regression

Dependent Variable = MF frequency of parent firms (or maybe frequency of combined firms)

Independent Variables:

* Information Asymmetry

* Spinoff Dummy (=0 if before spinoffs; =1 if after spinoffs) [expected to be (-)]

Control Variables:

* Total assets

* MB ratio (could also be a growth dummy, =1 if the firm is a high growth firm and its MB ratio is above the median of the set of all sample and control firms; =0 otherwise)



New Data from Prof. M

力力给了我他的Data,问有多少我的Database里没有。他有370个obs,有些70s年代就有。我的SDC始于1981年。370中,有一半,189个没有match。

例子:

29496 QUORUM HEALTH GROUP INC

http://www.nndb.com/company/224/000125846/

Quorum Health GroupCOMPANY

Spun off from HCA in 1991. It merged with Triad Hospitals, another HCA spinoff, in 2001.


有趣的是,这个obs在SDC Equity Database里面有。但是M&A里没有。

61574 WATERS CORP

In 1996, just after it went public following the spin-off, it reported having just over 2,000 workers. But when it was launched in 1958 by Jim Waters in the basement of the Framingham police station focused on building one-of-a-kind lab equipment, it had a total of six employees.

http://www.bizjournals.com/boston/print-edition/2013/09/13/waters-continuing-an-era-of-massive.html?page=all

也在SDC Equity里面。

13099 FRUIT OF THE LOOM LTD  -CL A

The company was part of Northwest Industries, Inc., until NWI was purchased by William F. Farley in 1985 and renamed Farley Industries, Inc.[1] Farley served as President, CEO, and majority shareholder for 15 years. 

http://en.wikipedia.org/wiki/Fruit_of_the_Loom

Mid-1980s Leveraged Buyout and Reorganization
Union did not escape the trend toward leveraged buyouts of the 1980s. In 1984, William F. Farley acquired Union Underwear when he bought Northwest Industries for $1.4 billion. Farley privatized the parent company and renamed it Farley Industries. In the 1980s tradition of leveraged buyouts and junk bonds, Farley parlayed his acquisitions into larger and larger conquests until, by the end of the decade, he had fashioned a textile and apparel conglomerate with $4 billion in annual sales and 65,000 employees worldwide.
In 1985, the conglomerate was restructured, $260 million in shares were sold, and Union Underwear was renamed Fruit of the Loom, Inc. to relate the business more closely to its famous trademark.
http://www.fundinguniverse.com/company-histories/fruit-of-the-loom-inc-history/
It was the third name change for the company, known as Northwest Industries until its acquisition in 1985 by Chicago financier William F. Farley in a $1 billion leveraged buyout. Farley, who heads Farley Industries, changed the name to Farley/Northwest Holding Corp. and later to Farley Apparel Inc.
The debt comprises $60 million of 6 3/4 percent convertible subordinated debentures, due March 1, 2002, and $250 million of 10 3/4 percent senior subordinated notes, due July 15, 1995.
The debentures are convertible at any time into Fruit of the Loom Class A common shares at $11.25 a share.
...
The company planned to go public last fall with a 30 million-share offering of Class A stock that was pitched largely to institutional investors. But apparently Wall Street balked at the terms: Underwriters had proposed to sell the stock between $14 and $17 a share.
http://articles.chicagotribune.com/1987-03-04/business/8701170677_1_convertible-debenture-offering-preferred-stock-loom

看上去不像spin off,也在SDC Equity里面。

7085 MASCO CORP
7086 METALDYNE CORP

  1. MASCO CORPORATION | Company profile from Hoover's ...

    www.hoovers.com/company/...CORPORATION/rfsjti-1-1njhxk....
    Hoover's
    Hoovers company profile of MASCO CORPORATION. ... In 1984 the firm split. ... wasspun off as Masco Industries, a separate public corporation (later Metaldyne ) ...
  2. 所以,以上是同一个obs.也在SDC Equity里面。
  3. 110065 INFINITY PROPERTY & CAS CORP
  4. A.M. Best Co. has assigned a financial strength rating of A (Excellent) to Infinity Property and Casualty Group, (Infinity P/C) the core auto subsidiaries of Infinity Property and Casualty Corporation (IPCC) [NASDAQ: IPCC] (Birmingham, AL). IPCC was created through a spin-off of American Financial Group, Inc.'s [NYSE: AFG] (Cincinnati) nonstandard automobile business produced through independent agents. The rating outlook is stable.
  5. http://www3.ambest.com/ambv/bestnews/presscontent.aspx?altsrc=23&refnum=7721
  6. 也在SDC Equity里面。