Robustness of productivity estimates
read more
Citations
Identification Properties of Recent Production Function Estimators
Creative Accounting or Creative Destruction? Firm-level Productivity Growth in Chinese Manufacturing
Cross-Country Differences in Productivity: The Role of Allocation and Selection
Exporting raises productivity in sub-Saharan African manufacturing firms
Identifying Agglomeration Spillovers: Evidence from Winners and Losers of Large Plant Openings
References
Measuring the efficiency of decision making units
Initial conditions and moment restrictions in dynamic panel data models
Technical change and the aggregate production function
Formulation and estimation of stochastic frontier production function models
Related Papers (5)
The Dynamics of Productivity in the Telecommunications Equipment Industry
Estimating production functions using inputs to control for unobservables
The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry Productivity
Frequently Asked Questions (10)
Q2. How does the DEA method estimate the contribution of plants to aggregate productivity?
Summing up changes in input shares, weighted by initial productivity, indicates that less productive plants used an even larger share of inputs by the end of the sample, lowering aggregate productivity by 21.1%.
Q3. What is the way to separate the productivity component from the random error?
If one observes firms only once, making strong assumptions is the only possibilityto separate the productivity component from the random error.
Q4. What is the common reason why the index numbers are hard-hit?
The index numbers are especially hard-hit when the error is proportional to the variation in the underlying variable without measurement error, not surprising given that the wage bill is one of the most volatile variables.
Q5. What is the main drawback of the normality assumption on wages?
The normality assumption on wages is convenient to obtain an explicit functional form for investment, but it can lead to negative wages.
Q6. Why did the truncated distribution for productivity have little impact on the results?
Because the31Incorporating a truncated distribution for productivity, as the stochastic frontier literature assumes, had little impact on the results.
Q7. What is the method for estimating productivity levels?
In the benchmark case, the Olley-Pakes method estimates productivity levels mostaccurately, especially if the random measurement error is taken out (OP2).
Q8. What is the conditional expectation of qit llit?
From the production function (10), one can write the conditional expectation of qit − αllit as α0 + αkkit plus the conditional expectation of productivity.
Q9. How do parametric methods perform when the errors get very large?
While they achieve an astounding correlation of 0.95 and 0.89 without measurement error, these fall by at least three quarters when the errors get very large.
Q10. How does the worst method achieve a correlation coefficient over 0.32?
No method achieves a Spearman rank correlation coefficient over 0.68, while the worst method still achieves a correlation of 0.32.