Trade in Tasks and the Organization of Firms. Dalia Marin, Jan Schymik, Alexander Tarasov University of Munich CESifo Global Economy Conference Munich May 2014 . The Changing Nature of Organizations and International Trade.
Trade in Tasks and the Organization of Firms
Dalia Marin, Jan Schymik, Alexander Tarasov
CESifo Global Economy Conference
Munich May 2014
(Rajan and Wulf (2006), Marin (2008), Marin andVerdier (2010), Bloom, Sadun, van Reenen(2010)
What is the connection between
trade in tasks,
decentralized management and
the rise in CEO pay?
We incorporate trade in tasks of GRH (2008)
small open economy version of the theory of firm organization of MV (2012)
how offshoring affects decentralized management and CEO pay in firms
we show that offshoring of production tasks
we show that offshoring of managerial tasks
We conducted a firms survey
among of 660 multinational firms in Austria and Germany with 2200 affiliates in Eastern Europe during 1990 – 2001.
80 percent of German investments in EE
100 percent of Austrian investments in EE
X-sector, differentiated goods,
Y-sector, homogenous goods,
Two factors of production:
human capital (skilled managers)
labor (production workers)
In the X-sector firms choose a firm organization
P has formal authority in the firm
P runs the firm with A’s cooperation
P delegates formal authority to A
A runs the firm with P’s cooperation
P runs the firm without A’s cooperation
We consider a firm with a simple hierarchy
CEO (P) hires a skilled manager (A) to run the firm and workers to produce, P and A look for projects for the firm
There is a conflict of interest between P and A
Payoffs of P and A depend on who’s project is implemented
P supervises the more
the larger are her stakes (the larger are
A has more initiative
the lower P’s probability of intervention
Cost of hierarchies: loss of initiative
From Marin and Verdier (2012) we get
Firms will choose decentralized management (A-organization) at an intermediate level of profits
At low and high profits there is no trade-off between control and initiative, hence, firms choose control (centralized management)
small open economywith
X and Y are
In a small open economy (SOE): thenumberandprices
Prices in the SMO are
n*(pm)1-σ = IMlevelofimportpenetration, exogenous
Productionissector X involves a continuumoftasksofmeasure 1. Performingeachtaskrequiresc(i)unitsoflabor. Itis profitable to offshore taskj, j(0,1)if
Marginal costsofthe firm:
ZX withIXthenumberoftasksoffshored. IX = 0 ZX = 1
Profits of a firm i
Factordemanddepends on theequilibriumorganizationP,A,Oand on theorganizational mix offirms.
k=Afor intermediate profits k=Ofor high profits
Upward sloping HH curve:
B/w increases, excess demand for managers, q/w has to rise
q/w increases, few firms are looking for a manager, excess supply of managers, number of firms n has to rise which happens when B/w increases
Upward sloping EE curve (war for talent curve):
As B/w rises, firm entry, number of firms is fixed by H, competition for managers, q/w rises
How are changes in openness affecting the
How are changes in
offshoring of productiontasks affecting the level of decentralized management?
Two opposing effects:
Zx wc(i)/Zx profits rise (productivity effect)
Other domestic firms become more productive as well, revenues and profits (revenue effect)
In an open economy the productivity effect dominates the revenue effect and B/w rises
When profits rise, P starts to monitor more potentially destroying A’s initiative,
When profits rise sufficiently, the trade-off between control and initiative favors initiative, P delegates authority to the skilled manager, decentralized management
‘Who decides in your company over the
Please rank between 1 … 5’
1 taken at headquarters (CEO)
5 taken by the divisional manager
intra-firm imports from all subsidiaries in percent of parent sales
instrument for offshoring of production tasks:
‘standardized foreign input’
Prediction 1: In a cross section of firms in an economy open to trade, multinational firms will have more decentralized management, when they are offshoring more production tasks to low wage countries.
Offshoring firms are 33.4 percent
than non-offshoring firms
when we instrument offshoring with ‘standardized foreign input’.
Managers perform a continuum of tasks of measure 1 requiring one unit of managerial labor.
Foreign managers do not receive authority in the firm.
Fraction of managerial tasks offshored is exogenous and given by IS.
Itis profitable to offshore managerialtasksif
q > q*
q, q*managerwages at homeandabroad
q(1- IS) + q* IS costofmarketentry
Proposition 3:In the P-equilibrium, there exists a cutoff level of openness of the economy denoted by IM, such that for IM > IMp: B/w and q/w are increasing in Is; and for IM < IMp: B/w is declining in Is, while the impact of Is on q/w is ambiguous.
Three effects of a rise in Is:
war for talent effect: Is cost of market entry firm entry, fix H q/w and B/w
labor market effect: Is demand for skilled managers q/w
competition effect: more firms find a manager n competition B/w
depends on the level of openness IM
large openness IM > IMp: positive war for talent effect dominates the competition effect, profits rise with Is
small openness IM < IMp: competition effect dominates the war for talent effect, profits decline with Is
depends on the level of openness:
large openness IM > IMp: war for talent effect pushing up q/w dominates the labor demand effect pushing down q/w,as a result q/w
small openness IM < IMp: labor demand effect large (large shift in HH) but war for talent effect also large (steeper HH), effect on q/w
‘How many managers of your parent company have been sent to the affiliate firm?’
offshoring of managerial tasks = 1 – managers sent
in 57 percent of all direct investments at least one managers has been offshored
on average 2.63 managers are offshored per investment project
max 39 managers offshored
Total compensation per board member of the firm (Kienbaum management consultancy) relative to average wage of the firm (firm survey)
Prediction 2: In a cross section of firms in sectors sufficiently open to trade, multinational firms will have more decentralized management when they are offshoring managerial tasks to low wage countries.
An increase in the sample mean of the fraction of managers offshored (1.48) reduces the level of decentralized management by 3.1 percent but increases the level of decentralized management by 4 percent in industries with a level of openness above the 25th percentile of the openness distribution.
Prediction 3:In a cross section of firms exposed to international trade, multinational firms will pay their CEOs lower wages when they are offshoring managerial tasks to low wage countries and they will pay their CEOs higher wages when the number of firms in the domestic market increases.
One additional manager offshored lowers CEO pay relative to workers by 6.9 percent (labor market effect).
This implies that relative CEO compensation is lower by 13 – 18 percent due to managerial offshoring.
One additional manager offshored allows one additional firm to enter which increases relative CEO pay by 2 percent (war for talent effect).