Transcript Turnover-based productivity growth

Turnover-Based Productivity
Growth
October 31, 2005
James Tybout
Pennsylvania State University and NBER
Productivity decompositions
• In standard productivity decompositions, turnover-based
productivity growth (TBPG) occurs when:
– Unproductive firms exit and more productive firms enter.
– Market shares are reallocated from low productivity firms to high
productivity firms.
• The other source of productivity growth is intra-firm
productivity growth (IFPG).
–
–
–
–
Vintage effects
Endogenous innovation
Learning spillovers
Random shocks
TBPG and IFPG are jointly
endogenous processes
• Because of vintage effects and endogenous innovation,
IFPG reacts to turnover.
• TBPG, in turn, reflects the patterns of decay in relative
productivity among incumbent firms.
• Transition dynamics can be subtle, and time horizons
can be lengthy.
Descriptive analysis of TBPG has
its limits
• Sometimes major policy shocks make it possible to draw
inferences from descriptive statistics (e.g., Bartelsman et
al, 2004, concerning the transition economies).
• But often, it is difficult to assess an industrial sector’s
performance and its relation to policy using productivity
decompositions and reduced-form regressions.
• The alternative: structural models of industrial evolution.
Structural models with TBPG
• Because they are forward-looking models with
uncertainty and multi-agent optimization, industrial
evolution models are computationally cumbersome.
• The scope for variation in model specification is
extensive:
–
–
–
–
–
–
Endogenous versus exogenous innovation
Competitive versus imperfectly competitive market structures
Vintage effects versus homogeneous capital
Perfect versus imperfect capital markets
Partial versus general equilibrium
Open versus closed economy
• But progress is being made, so specific stories about
TBPG in specific contexts are starting to emerge.
• A (non-representative) sampling of developing country
studies follows.
Macro shocks and subsidies for
incumbent firms
• Bergoeing, Loayza and Repetto (2004) assume:
– Firm-specific vintage capital and exogenous productivity shocks
– endogenous turnover
• They explore the effects of production subsidies on the
speed of recovery from a negative aggregate shock.
• The subsidy keeps low-productivity firms active. So:
– subsidizing incumbents in the aftermath of a shock “reduces
volatility and firm destruction at the cost of a long period of
stagnation.”
– The loss in present value of GDP is a factor of 1.5 to 2.5,
depending upon the magnitude of the subsidy (3% or 6%).
– Qualitatively, the simulations are consistent with cross-country
correlations of regulatory indices with indices of the severity of
recessions.
Exchange rate volatility and
selection
• Pratap and Urrutia (2004) study industrial evolution in a
context with:
–
–
–
–
foreign currency financed investment,
exports,
potential bankruptcy
exogenous, idiosyncratic productivity shocks
• Main findings:
– Depreciation discourages investment by reducing the net worth
of firms with dollar-denominated debt, and thereby raising
borrowing costs.
– This effect can dominate the demand-side effect of depreciation
Exchange rate volatility, interest
volatility and TBPG
• Bond, Tybout and Utar (2005) assume:
– Firms are closely held by risk-averse households with
heterogeneous wealth.
– Firms’ profits are determined by their capital stocks, productivity
shocks, and the exchange rate.
– Households choose whether to create (or shut down)
proprietorships. Those that operate proprietorships choose how
much to invest in them, subject to collateral constraints.
Exchange rate volatility, interest
volatility and TBPG
• BTU (2005) estimate profit function and macro
processes using Colombian data.
• They explore the effects of crisis-prone macro
environments on TBPG and investment behavior.
• Key effects:
– Households with modest wealth less likely to operate
proprietorships during volatile periods.
– Incumbents with big, poorly performing firms that would have
exited in a stable environment are more inclined to hang on.
Simulated Transition to High Volatility
• Initially, volatility increases number of firms, relative to the base case
Simulated Transition to High Volatility
• The association between size and profitability is initially weakened
by volatility—big, poorly performing firms are induced to hang
around
Simulated Transition to High Volatility
• The poorly performing firms that hang on reduce size-weighted
productivity significantly.
Import competition and TBPG
• Erdem and Tybout (2004) examine the effects of import
competition using an open economy version of the
Pakes and McGuire (1994) model:
–
–
–
–
–
Oligopolistic market structure
Endogenous investments in innovation
Vintage capital effects
Imports are imperfect substitutes for domestic varieties
Quality of imports increases stochastically
• ET (2004) study the effects on TBPG (and welfare) of a
reduction in the price of imported goods; accelerated
innovation abroad.
Reduced Import Prices and efficiency
Unweighted Mean Efficiency, RPM Experiment
4.9
4.8
4.7
4.6
4.5
4.4
rho=0 after t=50
4.3
rho=1.5
4.2
4.1
100
91
82
73
64
55
46
37
28
19
10
1
4
time
Average quality of the domestic goods initially improves relative to
imports.
Why? The worst firms/product lines immediately shut down, and
the deterioration due to reduced investment is gradual.
Reduced import prices and
innovation
Unweighted Mean Investment, RPM Experiment
0.9
0.8
0.7
0.6
0.5
rho=0 after t=50
0.4
rho=1.5
0.3
0.2
0.1
100
91
82
73
64
55
46
37
28
19
10
1
0
time
But incentives to innovate are less because profits are
lower (Schumpeter).
So firms have shorter average life expectancies (not
shown), and entry/exit go up.
Reduced import prices and producer
surplus
Producer Surplus, RPM Experiment
2.5
2
1.5
rho=0 after
t=50
rho=1.5
1
0.5
100
91
82
73
64
55
46
37
28
19
10
1
0
Producers earn less surplus in the new regime:
–smaller mark-ups,
–smaller market shares,
–shorter life spans.
Accelerated improvement in imported
goods
Weighted Mean Efficiency, AIM Experiment
16
14
12
10
8
6
4
2
0
delta=.6
97
89
81
73
65
57
49
41
33
25
17
9
1
delta=.8 after
t=50, relative to
counterfactual
time
Domestic goods quality improves rapidly because new
firms/product lines embody best practice.
Turnover is more rapid because firms have more incentive to be
new.
Accelerated improvement in imported
goods
Unweighted Mean Investment, AIM Experiment
97
89
81
73
65
57
49
41
33
25
17
9
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
delta=.8 after t=50
time
delta=.6
But heightened import competition reduces producers’ incentives
to invest in innovation (Schumpeter wins again).
Concluding remarks
• Theoretical models provide some interpretations for
TBPG and IFPG, and their relation to policy.
• But the relationships are subtle, dynamic and very
dependent upon specific modeling assumptions.
• Some findings particular to special contexts have begun
to emerge, but much remains to be done.
Concluding remarks
• Some directions to push in:
–
–
–
–
–
Move toward econometric estimation
Allow for endogenous innovation.
Model entrants in a more realistic way
Do a better job of characterizing performance
Study entry costs more closely—when are they excessive?