Transcript Firm-Level Implications of Demand Fluctuations

Firm-Level Implications of
Demand Fluctuations
Using Regression Results in Decision Making:
Advertising Mix, Interest Subsidies and
Pricing Policies
Forecasting Demand and Inventory Management
Business Cycle Aspects
1
Choosing Price and Advertising
• QD = B0 (Price)b1 (ADV)b2
• LogQD = b0 + b1 LogPRICE + b2 LogADV
• Here b1 and b2 are elasticities.
• Reasonable to expect: b1 < -1 , 0 < b2 < 1 :
2
Elasticities
LogQD = b0 + b1 LogPRICE + b2 LogADV
If Price rises by 1% then QD will rise by b1%
Clearly b1 < 0. Say b1 = - 0.5: firm will keep raising
price forever -> price rises by 1%, quantity falls by
0.5%, so revenue rises, but costs fall. Need b1 < -1.
Clearly b2 > 0 . Say b2 > 1: ADV up by 1% raises
demand by more than 1%. If marginal production costs
are not increasing, the firm will keep increasing ADV.
3
So 0 < b2 < 1 reasonable.
Profit Maximization
Firm chooses PRICE and ADV to Max. Profit.
Profit = Revenue - Cost
Revenue = PRICE x QD
Cost = CAVC x QD + FC + PRADV x ADV
CAVC = constant average variable cost, FC =
fixed cost, PRADV = price of advertising.
Note that we ignore inventories and that all
demand is assumed to lead to sales.
4
Optimal Price
PRICE = [b1 / (1+b1)] x CAVC
Price is set as markup above average variable cost.
Say b1 = - 2 in regression, then Optimal Price = 2
times CACV
The more competitive the industry the more
negative will be b1. In perfect competition
b1 = - infinity ->
PRICE = CACV
5
Advertising Mix
LogQD = b0 + b1LogPRICE + b2LogADVNP +
b3LogADVTV
Cost: PRICENP x ADVNP + PRICETV x ADVTV
Say you adjust advertising mix by raising NP and
lowering TV advertising, keeping cost constant:
PRICENP x  ADVNP + PRICETV x  ADVTV = 0
 ADVNP /  ADVTP = - PRICETV / PRICENP
When is the resulting % change in QD positive? :
b2 %  ADVNP > - b3 %  ADVTV, or:
6
b2 ( ADVNP/ADVNP) > - b3 ( ADVTV/ADVTV)
Since ADVNP/ADVTV = - PRICETV / PRICENP
ADVTV/ADVNP > (b3/ b2) (PRICENP/PRICETV).
Thus, if the above holds, you will increase QD
keeping costs the same as long as you raise ADVNP
while lowering ADVTV. You should keep doing so
until you no longer benefit. At that point:
ADVTV/ADVNP = (b3 / b2) (PRICENP/PRICETV)
7
Example Problem
LogQD = 20 - 3 LogPRICE +
0.4 LogADVNP + 0.8 LogADVTV
FC = 10, CAVC = 12, PRICENP = 3 per ad,
PRICETV = 30 per ad.
Find the optimal Price and Marketing Mix:
PRICE = 1.5 x 12 = 18
ADVTV/ADVNP = (0.8 / 0.4) (3/30) = 0.20
8
Rebate or Interest Subsidy?
QD = a0 + a1 PRICE + a2 REBATE + a3 ISUB
Expect: a3 > 0 , a1  - a2 (?)
Cost: [REBATE + ISUB (PRICE-REBATEDOWNPAY)] x QD
Increase rebate while lowering interest subsidy,
keeping QD constant:
 REBATE /  ISUB = - a3 / a2
9
 REBATE /  ISUB = - a3 / a2
What is the change in Cost?
[REBATE + ISUB (PRICE-REBATEDOWNPAY)] x QD
 Cost = [ REBATE (1 - ISUB) +  ISUB
(PRICE-REBATE - DOWNPAY)]QD < 0 ?
Using the equation at the top of the page:
PRICE - REBATE - DOWNPAY >
(a3 / a2)(1 - ISUB) ?
10
Thus, you can lower your cost by
increasing your Rebate and decreasing your
Interest Subsidy, leaving demand
unchanged, until the equation holds with
equality:
PRICE - REBATE - DOWNPAY =
(a3 / a2)(1 - ISUB)
11
We have looked at the use of regression
results for making Pricing and Marketing
Mix decisions.
Next we will consider how regression
results can be used to forecast demand
which will help make Production and
Inventory decisions
12
Motives for Holding Inventory
• Bufferstock -- if costs of stocking out
exceed inventory holding costs. [for
materials, if supply interruptions may
occur]
• Production Smoothing -- often a
relatively stable level of production
is less costly, so in times of low
demand you may be best off
producing for inventory to stabilize
production. [for materials, transport
may be cheaper]
13
Cost
C(Q2)
Production
Smoothing
C(Q)
C(Q)
C(Q1)
Production
Q1
Q1 + Q2
2
Q2
14
More motives for holding inventory:
• Production Bunching -- in industries with
high start-up you may be best off
producing high volume (exceeding
current demand) when you do produce.
[for materials, if ordering costs are high]
• Speculation -- holding inventory because you
expect your product to appreciate in
value. [for materials, you may expect
suppliers to raise prices].
15
Costs of Holding Inventory
• Storage (warehouse space, surveillance,
upkeep, transportation)
• Interest foregone (real interest only)
• Insurance (or risk disaster: fire, flood, etc)
• Theft (employees or others)
• Depreciation (nature or fashion)
16
Value of Forecasting Demand
• Better forecasts reduce demand
uncertainty, decreasing the need for
bufferstocks.
• Forecasts of future demand help firms
smooth production, lowering average
production costs.
17
Automobile Industry and Inventory
Production Smoothing and Bufferstock
motives for finished goods.
Bufferstock and Bunching motives for
materials.
Significant: storage costs, interest loss,
depreciation.
18
Trends:
Recently, decreases in inventory-to-sales
ratios. More so in intermediate stages of
production than in final goods.
Reasons: improvements in transportation
(deregulation, competition) and
communication, bar-coding and
computerization.
Just-In-Time inventory policy: fashionable,
lowers inventory costs but exposes firms to
interruptions, forecasting becomes crucial. 19
Inventories and the Business Cycle
Invt+1 = Invt + Prodt - Salest
Invt+1 = Invt + (Prodt - DemandForecastt)
- DemandShockt
Invt+1 - Invt = PlannedInvt + UnplannedInvt
Planned inventory increases signal a business cyle
upturn, Unplanned increases signal a downturn.
Inventory-to-sales ratios are procyclical in U.S. data,
suggesting most inventory increases are planned.
20
• Unplanned inventory increases cause firms to
produce less in the future, which lowers future
incomes and future demand. As a result, one-time
negative demand shocks have persistent effects on
the economy. These effects are sometimes described
as inventory cycles.
• Inventory changes during the business cycle are on
average larger than changes in GDP over the cycle
(Table 1, Allen article).
• Reductions in inventory-to-sales ratios have not
decreased the amplitude of business cycles.
21
Differences in Manufacturing vs.
Services Demand Volatility
• No pent-up demand for services. Manufactured
goods purchases may be deferred until one can better
afford them, accentuating recessions.
• No inventory build-up in service sector. So less
susceptible to inventory cycles.
• Few services are tradable internationally, so not
subject to exchange rate fluctuations
• Services typically do not require much capital so not
subject to interest rate fluctuations.
22
Chart 2 in Filardo article shows the dramatic
differences in employment variability between the
service sector and the recession-prone manufacturing
sector.
The shift over time in employment from
manufacturing to services (shown in Chart 1),
however, has had little effect on the amplitude of the
business cycle (as shown in Chart 3). Reasons: the
output share of the manufacturing sector has barely
changed; outsources by manufacturing firms now
counts as services what used to be counted as
manufacturing; services have become more cyclical.
23