Transactions, Incomplete Contracts, Coordination, Vertical Relations,

Information Economics

State, Signal, Action

Information signal provides some valuable information about the true state, and decision maker benefits

Basic Assumption: No objective misalignment in decision making

Principal-Agent Problem

Principal is profit maximizing by choosing a compensation scheme for the agent

Subject to agent paid better than outside option and agent maximizes payback by choice of effort

Another way - 2 stage game: Princ chooses S, agent chooses a, outcomes are observed.

Findings

No IC Condition - choose best action ST agent paid reservation wage

When principal is risk neutral, numerator is 1, so compensation is a constant. Principal takes all risk, agent - no risk. y (other performance information) adds no value.

G'(x-S(x,y))/U'(S(x,y) = Lambda

IC condition added

Term gets added to the Lambda. A fixed compensation will make the agent shirk, compensation needs to be based on x and y as well.

y needs to be information in addition to x to be useful

y needs to be something that agent effort 'a' can affect

Revelation Principle

For any non-truthful reporting strategy, principal can design a contract that will induce more truth telling

Coordination

Specific vs General knowledge

Time, location specific

granular, poor aggregation properties

Dispersed

Continual change, dontinual decisions

Passing required information downward

Prices

Only the relative extent of unavailability, not why, etc

An inducement to use less of the product in shortfall.

The Problem

How best to make decisions through economic interactions of people when there is partial information

Solution 1: Collocation of decision making rights with knowledge important to these decisions

Problem: What if there is no alienability? Like intra-firm

Solution: Hayek says decisions rights will move to agents with knowledge. Moving up or down depends on COSTS of moving knowledge

The intra-firm Principal-Agent Problem

Causes: State not observable, action not calculable, objectives misaligned,

Effects of IT on centralization vs decentralization: Ease of specific knowledge transfer to CEO; improvements in control technology

Contralization vs Decentralization

Consider coordination between HQ and nodes

Problems

Information Asymmetry

Nodes know own demand (Anand &M 1997, Nault 1998, Nault & Deter 1994, Nault and Dexter 2006,), or own productivity (Harris et al)

General Setup1: There are investment complementarities between HQ and nodes. But each doesnt know the other's investment level. CA sets m,t. Org forms include Hierarchical, decentralized, mixed (Nault papers)

Externalities

Node's investment increases other nodes profits and HQ profits, leads to under-investment(Nault 1998, Nault and Dexter 1994, Nault and Dexter 2006)

What should investment levels be for nodes and HQ?

Nodes do not know HQ's production cost (A&M1997)

HQ does not know node's productivity

General Setup 2: Nodes place orders with HQ, or report productivity with HQ (A&M1997, Harris et al 1982)

Findings

IS reduces coordination cost

Collocation can be good. But if info can be moved through IS, decisions can be made at either end

Using revelation theorem, nodes can be made to reveal true productivity

t=0 is ideal, to reduce horizontal externality

Sharing of information between supply chain partners can increase efficiency of supply chain

Inter-firm coordination : Boundary of the firm, transaction costs, incomplete contracts and asset ownership

Transactions

Behavioral assumptions

Bounded Rationality

Opportunism

Uncertainty

Frequency

Asset Speficity

Location

Physical asset

Human Asset

For market (buying), as specificity increases, governance cost increases relative to making. Production costs are assumed to always be lower in the market. Net, making starts to look better

Cooperative Game Theory

Problem: To find a function that will divide value added to members

Axioms

Efficiency: Allocate Everything

Dummy: Add nothing, gain nothing

Symmetry: Labels do not matter

Function should be additive

Shapley Value

Sum over all allocations ((prob of being in a coalition)*(value added to that coalition)

This is one of the characteristic functions that will satisfy the above axioms

Non-Cooperative Bargaining Solution

Here, axioms lead us to the solution: Split the difference!

Problem: Given incomplete contracts, how should we allocate asset ownership to achieve high investment levels

Incomplete Contracts

Stem from bounded rationality, some things cannot be fully contracted: care, creativity, quality

In such cases, uncontracted quasi-rents (Residual rights )go to property owner

An agent with no access to assets risks no return on investments on HR skills. The risk can be mitigated if he is assured access to the asset inherent in ownership.

Core: the best set of allocations that cannot be improved by reducing the size of the coalition

Solution and Findings

Integration can reduce holdup problems.

Agents under-invest if they find that some of the returns from their investment will dissipate

Asset ownership gives the right to exclude others from using it

General Setup

Stage 1: Agents invest non-cooperatively Xi; Stage 2: Agent value is realized based on Shapley value.

Stage 1 is value maximizing non-cooperative investment. Put your money where the best return is, much like a stock market.
Stage 2: Ex-post value generation is visible, greater the value generated for the coalition, greater the return - an axiom. Presence of industry bodies and standards even in highly competitive industries exemplifies this. X's are observable at date 1 to all.

FB: FOC is when marginal benefit of each investment is equal to cost across all agents

But in reality, at date 0, each agent i chooses Xi to maximize Bi(a|x)-Ci(Xi). The benefit is highly reduced because of uncertainty

If only one agent has investment, he should own all assets

Not more than 1 agent should have veto power over an asset

If an agent is indispensable to an asset, he should own it

Strictly complementary assets should be owned together

If an asset essential to an agent i is controlled by agent j, then agent j should control all assets

Bakos and Nault 1998

RQ: Derive optimal ownership structure for the internet

Vi(S,A|X) = Sum(LAMinAn)Sum(MUikXk)Xi^1/2

IT reduces in-house production costs as well as transaction costs, favoring both markets and hierarchies, net impact is inconclusive

They explore through a functional form, how H&M1990 can be applied to IT networks. LAMin scales the impact of asset An on value realized by i. MUi scales the impact of each agent's investment Xk on i's value generated.

Methodology

FOC: FB: Marginal Benefit = Marginal Cost

FOC: Marginal increment in Shapley value = Marginal Cost

Comparing the above, there is underinvestment

Brynjolfsson 1994

If agent 2 has asset ownership

Agent 1: 1/2(MB) = MC

Agent 2: MB = MC

Results: Asset Specificity in supply contracts without asset ownership are prone to under-investment

Bhardwaj et Al 2007

Research Question

Does IT synchronize externally oriented functions like marketing and Supply chain, with internally oriented functions like Manufacturing

Does IT enhance manufacturing's coordination with supply chain and marketing in order to increase manufacturing's performance

Specification

3 Direct effects, interactions, controls and Inverse Mills Ratio

Yao and Zhu 2012

Research Question

Do electronic linkages reduce the bullwhip effect?

Defn: Bullwhip effect is defined as the amplification of demand variability from a downstream site to an upsteam site

Findings

EL use with supplying industries tends to reduce bullwhip effect

EL use with buying industries tend to increase bullwhip effect

IT seems to mitigate this effect

Hypothesis justification: IT reduces coordination costs : Cost of exchange and processing information. 2) Monitoring costs 3) Increasingly standardized, reducing asset specificity 4) Order batching

Anomalies

Buyers have more options when using EL, they can switch, causing variance in demand

Defn: A Transaction occurs when goods or services are transferred across a technologically separate interface