Monday, February 28, 2011

Some Observations on the Nature of Contract Drafting

It is time to be more controversial. I’ll start with my favorite quote on contract drafting written by one of the greatest jurists in modern times. Lord Denning, in his Romanes Lecture, From Precedent to Precedent, given at Oxford in 1959, said:

“[Lawyers] will so often stick to the letter and miss the substance. The reason is plain enough. Most of them spend their working lives drafting some kind of document or another – trying to see whether it covers this contingency or that. They dwell upon words until they become mere precisians in the use of them. They would rather be accurate than be clear. They would sooner be long than short. They seek to avoid two meanings, and end – on occasions – by having no meaning. And the worst of it all is that they claim to be the masters of the subject. The meaning of words, they say, is a matter of law for them and not a matter for the ordinary man.”

I submit the substance of drafting lies is in the configuration of the agreement and its clauses as much as the precise choice of words. Having spent a number of years working on the empirical analysis of contracts, I can share a number of observations.

1. There are just a handful of contract clauses that are critical

These are the terms that define a transaction—the highly negotiated clauses that the parties seek to negotiate in their favor. Profusek and Ganske state the matter plainly: “Many lawyers no longer add real value to dealmaking. That's perhaps a startling and somewhat harsh allegation - especially from a couple of M&A lawyers - but we're afraid it's true.” It's Time To Rethink The Lawyer's Role In Dealmaking: Start By Facing Up To The New Realities, Metropolitan Corporate Counsel, December 1, 2007. The Jones Day partners assert that: “Half of the words merely repeat what has been said somewhere else. Really now, has anything bad ever happened because all the Form 5500s weren't filed? Half of the remaining paperwork is boilerplate, leaving only a handful of provisions that are important: the money, closing, social, and fiduciary provisions.” Supplementing this practitioner’s viewpoint, I can add that technology will soon be able to detect the valued terms, and by deduction identify the less important provisions.

2. All contract clauses have a standard set of terms

Contract analysis can find the core, pre-negotiated language by indentifying the common terms and distinguishing deal-specific or negotiated terms. Applying this technology, it can be seen that differences between individual clauses are mostly semantic—the language may differ in clarity, format and voice—but any lawyer or judge will recognize its intent, no matter how imperfectly or inarticulately drafted. Of course, there does come a point when a clause is so poorly drafted that—even if you know what it is trying to say—it cannot be given is normal interpretation, because it is excessively vague or open to too many different interpretations.

3. The clause standard is not party neutral

Standard clause language is not evenly balanced between the interests of the parties. Clause weighting in favor of one party or another varies with each agreement type. For example, the ISDA master document is evenly balanced. An employment agreement is typically weighted in favor of the employer; while an end user software license agreement is heavily weighted in favor of the licensor. For example, Florencia Marotta-Wurgler analyzed 647 license agreements, plotted their bias on a standard distribution curve, and found that the agreements are typically weighted in favor of the sellor. What's in a Standard Form Contract? An Empirical Analysis of Software License Agreements, New York University Law and Economics Working Papers, July 29, 2008.

4. Clause variations add or subtract language elements

Clause variations shift the benefits and burdens in favor of one party or another by adding or subtracting terms from the standard. For example today’s clause of the day on theContractStandards web site illustrates the options in a non-competition clause.

5. There are relatively few substantive variations for each clause

While there are an infinite range of semantic variations, there are typically less than five substantive variations within each clause.

Conclusions

If these observations are correct, then we can draw a couple of conclusions.

1. There is not a single, optimal clause. Rather, all clause examples represent a point on a continuum from being favorable to one party or another. While is helpful to propose optimal clause language, it is more useful to understand what is market standard and what are the optional variations.

2. Document assembly may not be the best approach to automating document production, because it is very unlikely that anyone can develop a set of logical rules in advance to set the appropriate language modifications. The final document is more likely the product of a negotiation.

Monday, February 14, 2011

New Web Site--Contract Standards

KIIAC LLC (KIIAC) announces the launch of ContractStandards.com, a web site that seeks to collect and share global contract standards, document automation tools, and provide a framework for collaborative development. The site’s mission is to openly share contract standards—transaction analysis, checklists, and clause libraries—in an effort to establish global contract norms. It is hoped readers will share their knowledge through comments, practice guidance and other contributions.

Contract Standards will also develop and share document assembly tools. The initial suite of tools is provided by KIIAC under a General Public License. All participants are invited to suggest new features, propose support for different platforms, or enter into a discussion of document automation in the legal profession. It is further hoped that some individuals and organizations may contribute development resources to further refine the tool set.

The site will feature a “clause-of-the-day” showing the standard clause language determined by statistical analysis of documents filed on EDGAR, together with a range of clause variations.

Please visit the site. Better still, contribute.

Friday, February 11, 2011

Step 7. Markup—Model Document for Assembly

Task: Prepare document for assembly by replacing client-specific information with text variables and options that can be automatically configured to client-specific needs.


Once the form has been reviewed and edited, it may then be marked up for automated document assembly. Some firms may automate the document using commercial software, such as HotDocs, DealBuilder, Exari or BrightLeaf; others may utilize a Microsoft Word macro to quickly conform a document to specific client needs.

The automation steps may include:
  • Links to online clause libraries
  • Links to online practice guides deliveries through a Wiki
  • Tools to automate selection of alternate clauses and optional language with Word macro
For those firms that have not developed a macro, kiiac will release a Word automation tool into the public domain in late February. This macro has the capacity to handle text variables (names, dates and amounts, etc.) optional text and questions. It is hoped that by releasing the code under a General Public License, readers will propose new features, and some firms or developers may assist with further development.

When the macro is released, a link will be added to this blo post.

Step 6. Edit—New Form Document

Task: Review and edit clause language to ensure best practices.

Lawyers should review the new form document to ensure all clauses conform to best practices. This step is, of course, time-consuming and exacting. However, techniques are available to reduce the time required and ensure quality.



  • Source: Where the new form—created in step 5—is an existing template, perhaps created in an earlier effort at standardization, it is possible to mark the source of each clause, allowing the reviewing attorneys to focus their time on the alternative clauses drawn from different source documents and spend less time on the previously reviewed template clauses.


  • Clause Standardization: One of the most effective methods to ensure high quality forms developed in the most efficient manner possible is to standardize clause language across different document types. Standard clause sets can be built for the Representations, Warranties, Covenants, Conditions and Miscellaneous sections of the agreement. In some cases, different clause variants must be used. For example, the assignment clause may permit transfer of contractual rights in the purchase of commodity assets, but will not permit assignment of a personal service contract. This modular approach is effective because there are a relatively small number of common clauses and a small number of common variations.

Once the document is edited, it can be added to the template, analyzed by the software and all its clauses marked as the new default clauses and alternative provisions.


Review Tools for Language Review



  • Online cluase editor tools

  • Microsoft Word document containing most conforming document clauses, and where divergent, the most conforming clause (from all documents) as an alternative clause

Thursday, February 10, 2011

Step 5. Generate—Base Document and Clause Alternates

(a) Select Base Document(s)

Task: Identify the base document or documents upon which to build a standard form.

A highly conforming base document is typically selected as a starting point for a new form because all of its clauses and definitions will have been drafted to work together, thereby eliminating additional work to conform language if the clauses were selected from different source documents. This document is also analyzed to assess whether it also represents a de facto standard, based on whether it is widely used by different lawyers and offices (i.e. there are a number of similarly constructed documents).

If a de facto standard base document is not found (which is rare), further research may be conducted to determine if there are multiple base documents (perhaps used for different types of transactions or in different jurisdictions). Alternatively, the review team may determine to standardized on one source document, not necessarily the most conforming, or to standardize on a published form (such as an ABA model form).

Review Tools for Document Organization
Online document review (select Source Document List from View menu):

(b) Choose Clause Alternates

Task: Determine the standard and alternative clauses that will available in the new form.

  • Missing Clauses: Identify the clauses in the most conforming document that might be missing from the standard established by all source documents. These are clauses that occur frequently in the document set, but are absent in the most conforming document. Determine whether to add the most conforming clause from the clause library or different clause example, if desired.
  • Divergent Clauses: Identify the clauses in the most conforming document that diverge from the standard core language established by all source documents. These are the clauses that may be missing some standard language or may contain deal-specific language. Determine whether to replace these clauses with the most conforming clauses or another clause example, if desired.
  • Weighted Clauses: Optionally identify additional clauses from documents drafted in favor of one party. This process can be quickly performed by finding “marker” clauses indicating pro-buyer or pro-seller situations. For example, an “as-is” warranty will likely reflect a pro-seller document. All clauses in this document can then be tagged as pro-seller and selected for output in the initial form document.

Review Tools for Clause Selection

  • Online document assembly template
  • Microsoft Word document containing most conforming document clauses and the selected alternative clauses, displaying information why such clause was selected (e.g. pro-buyer)

(c) Assemble Initial Form Document

Task: Create an initial form document for attorney editing, containing the clauses from the base document, missing clauses, and alternate clauses.

Select clauses for inclusion in the initial form and assemble the document. By default, the document will contain embedded links to the clause libraries.


Step 4. Review—Template

Task: Review document outline with lawyer-sponsor and confirm it is appropriately organized.
  • Is it organized appropriately?
    • How many levels should the outline contain?
    • Should the outline contain all clause types, or just the most common?
    • Should uncommon clauses be grouped into one bucket?
  • Does it combine appropriate clauses together?
  • Does it contain any duplicate provisions?
  • Is it organized in the correct sequence? (lawyer)
  • Does it contain all required provisions? (lawyer)
Sample Issue 1: the source documents variously contain sections titled Covenants, Covenants of the Buyer, Covenants of the Seller, Additional Agreements, etc. Should these sections be combined into a single provision titled Covenants? The answer will be found by determining whether similar clauses appear in each section. If the answer is yes, the sections are overlapping and may be combined into a single section.

Sample Issue 2: the source documents contain numerous variations of a Further Assurances clause, sometimes found in the Sale and Purchase, Covenants and Miscellaneous sections. By examining the language, it can be determined if these clauses are duplicates, and, if so, the clause frequency statistics can be used to determine the location of a single instance of the clause in the template outline.

Review Tools for Document Organization and Quality Control

(a) Outline (Document Checklist): Available online by selecting Outline from the tab menu or through an Excel report. These tools are used to review document organization and determine whether each outline provision is constructed from the correct building blocks. This review is performed by examining the alternate captions or, in other words, the names of sections aggregated into a template section. For example, a Miscellaneous section may be built from sections in other documents titled General Provisions or Other Provisions.
  • Does the outline contain appropriate number of levels? Does a section have too few clause examples to warrant the addition of a deeper level?
  • Is the list of clauses in a section too long? Should clauses having 2 or 3 examples only be grouped into an ‘additional clauses’ category at the end of the section?
  • Are the clauses correctly combined? Review the list of alternative captions. Are they similar? If not mark the section for further lawyer review.
(b) Index (Clause List): Available online by selecting Index from the Tab menu or through an Excel report. These tools are used to review the clause list and determine whether each clause is captured in the correct outline section and to check for possible clause duplication.
  • Are there similar clause captions in the list?
  • Does a clause appear in two different sections of the document?
(c) Confidence: The Excel report also provides information regarding the program confidence that the software has identified a good set of clause matches. Following the caption, the report indicates the total of number of matching clauses. It also shows the number of clauses that are strong matches and the number of clauses that are weaker matches, because the clause has less text matches and/or is located in a different section of the document compared to the strongly matched clauses. This information can be used to determine which clauses section warrant further attorney review.

Step 3. Organize—Process Template


The organization step captures information about the document (such as title, parties, and jurisdiction), produces an outline of the document, and creates clause libraries for each section of the outline. All these steps can be performed manually, but they are much faster and more accurate when processed with technology.

Task: kiiac consultant processes the document collection, creates a template outline, and clause libraries.

Index documents:
  • Classify document type
  • Extract metadata (e.g. title, parties, law firms and jurisdictions)
  • Decompose into clauses
Generate template:
  • Create a template outline
  • Clause libraries
  • Generate review reports

(a) Metadata

Extract metadata from the source documents such as: title, parties, date, law firms and jurisdiction.

(b) Outline

Review all the documents and identify the common organizational theme. In some instances, the documents may be organized in a consistent manner. For example, all documents may have a section titled Miscellaneous or General Provisions. In other situations, similar clauses may appear in different sections of the outline. The program or reviewer must determine whether:
  • clause sections are distinct (such the organization clause that may appear in both the representations of the buyer and the seller);
  • clause sections are the same, but in different places in different documents (such as a further assurances clause that may appear in the transaction section, the covenants and the Miscellaneous provisions);
  • clause sections are overlapping (for example an amendment and waiver clause overlaps with both the separate definition of amendment and waiver clauses).
This step can be performed manually or automatically. Those who have attempted to organize a template manually will appreciate the complexity of the challenge as each organizational adjustment will have significant collateral effects. It truly is a job better performed by a computer, because technology can analyze millions of comparisons per second to find the best fit. The process is similar to the board game Mastermind; but played by a computer, which can solve the code in a split second.

(c) Clause Library
For each branch in the outline identify all the matching clauses. Within these clauses find the core language, namely, the clause that contains all the common terms and the less amount of divergent or deal-specific language. In addition, identify the full range of language variations.

See, How Does it Work for more information on the computerized process.

Step 2. Collect—Source Documents

Task: Choose the document type, options and capture source documents
  • What set of source document?
  • How can they be identified?
Question: What document sources should be used to build the template?
  • Internal documents (FileSite; West km)
  • Public documents (EDGAR)
  • Client documents
Question: How many documents are required?
  • Breadth: The number of documents should be sufficient to cover broad range of deal circumstances and jurisdictional differences, but not too many as to overwhelm the review process.
  • Consistency: The more consistent document set, then fewer will be required to build a template
  • Range: As a rule of thumb, a range of between 50-100 documents is optimum. Where the document set is highly consistent or where there is a narrow range of deal alternatives, then 20 documents or less can be used. On the other hand, where there is less consistency and more deal alternatives, then up to 250 documents may be required.
Question: Should the documents be vetted and reviewed before processing?

Where the process is performed manually, the documents will be reviewed as part of the process. However, if automation is applied, it is generally better not to perform vetting prior to analysis. While some may argue this risks the paradigm of "garbage-in; garbage-out", the issue is rather that most organizations will have limited tools to review document before processing; absent of reading all the documents. Once they are processed, the technology offers exacting tools to select the documents for inclusion in the template. As a result, most firms perform a simple document title search in their file systems or search platforms, sometimes limiting the scope by date and/or practice group.

Step 1. Plan—Select Document Type and Form Options

Task: Educate lawyer, gather requirements, and answers questions

Project scope and schedule are typically addressed in a kick-off meeting. In most cases, the meeting can be conducted by KM personnel or library staff using question and answer format to develop a project plan. The initial meeting should address the following questions.

  1. What document type?
  2. Is this a new form development project or a forms maintenance and update project?
  3. What is the source of the documents?
  4. How many documents will be used to construct the template?
  5. How will the template be used?

Question: Why Create Templates?

  • Quality: Standardized forms with alternatives clauses are important to creating high quality documents in an efficient manner. (See, http://www.kiiac.com/why.htm)
  • Profitability: In a highly competitive practice, some of these efficiencies can be passed onto the client in the form of lower fees; some can be retained by the firm in the form of higher margins. In addition, the use of forms and templates can increase profit margin when used in a fixed fee or AFA representation. (See, http://www.kiiac.com/financial.htm)

Question: Why not re-purpose documents from the last deal?

  • Deal Checklists: The last draft may not contain all the deal options appropriate for the needs of the current transaction.
  • Institutionalizing Poor Drafting Practices. Using the last draft brings forward the good, the bad and the ugly, potentially institutionalizing poor drafting practices. (Retooling Your Contract Process for the Downturn: Adams Drafting)
  • Failure to Capture Best Practices: Using the last draft means that new and improved drafting practices are captured in a particular set of deal documents, which may never resurface. A template approach, on the other hand, is a centralized repository to capture best practices and to weed-out poor drafting, thereby continuously improving the precedent collection.

Question: Why not use public forms freely available on the Internet or through subscription?

  • Real Practice: In many cases, public forms lack the detailed range of deal alternatives found in real practice. They are typically simplified, generic forms of a transaction.
  • Brand: Re-purposing a publicly available form risks commodity pricing, since any law firm can provide this service. Developing a firm’s drafting standards established its brand and may allow the firm to charge premium pricing.

Question: What types of agreement make good candidates for a template?

  • High Use/Volume: Document types that are frequently used in the firm.
  • High Value: Document types that have a high value
  • Client Interest: In demand by current and prospective clients

Question: How are templates used in practice?

  • Research: Templates can serve as an educational instruments to help lawyers quickly understand what terms are found in a particular deal type; as a research platform to find standard and alternative provisions; and as a negotiation leverage tool to assess market standards or a particular law firm’s drafting practices.
  • Model Forms: Templates can be used to prepare model forms, containing clause alternatives, serving as an excellent starting point for the next drafting project.
  • Document Assembly: Model forms can be automated to insert client and deal specific information to rapidly generate a first draft, customized to particular clients needs.
  • Document Review: Templates also serve as a reference standard to compare and benchmark any document to determine if it contains: all required provisions; deal-specific provisions; and potentially missing provisions.

Question: How much lawyer time does it take to create a template and model form?

  • Manual Review: We can estimate the time to thoroughly prepare a template using the formula: number of source documents multiplied by average document page length multiplied by the time required to read and review each page. If, for example, the project consists of 50 merger agreements, each averaging 75 pages, and requiring 5 minutes to review each page, then the project will require approximately 2 months to complete. Of course, the number of documents and time taken to read each page can be reduced, but this approach risks failing to identify standard practices and the full range of deal alternatives.
  • Automated Review: A standard desktop computer can review and analyze 30 pages per second, completing the review of the 50 merger agreements in 2 minutes. This analysis facilitates an expedited schedule to create model forms.
  • Timeline: See the Project Planning Schedule in Attachment 1

Question: Are they any special required that should be considered?

  • Special requirements (e.g. NDAs)
  • Data privacy rules