占位,更新。
看点:
1. 文章使用1998-2007企业数据,分析中国制造业的生产率。
2. 有很多数据处理的技巧:比如不同代码的衔接、通胀的处理等。
3. 文章从生产率的提高、企业进入和退出、企业间要素分配三个角度阐述了生产率的变量和分解情况。
文章工作量大,引用了大量的文献用于比较分析,可谓鸿篇巨制。
结论
(1) Drawing on anunbalanced panel of firms between 1998 and 2007, we find firm-level TFP growthof manufacturing firms averaging 2.85% for a gross output production functionand 7.96% for a value added production function.
(2) Over the fullsample period, our results identify net entry as thesource of more than two thirds of total productivity growth.
In all, wefind that TFP growth coming from improvements in continuing firms (theintensive margin of TFP growth) and through net entry (the extensive margin ofTFP growth) was the source ofover half of value added growth inmanufacturing over the 1998-2007 period. TFP’s contribution to labor productivity growth is even higher attwo-thirds. Aggregate TFPgrowth in Chinese manufacturing remains constrained by limited efficiency-enhancinginput reallocations between active firms, confirming results in Hsieh and Klenow(2009).
政策启发
As inputgrowth slows and the technology gap with advanced countries narrows, furtherreforms to enhance efficient allocation of resources still provide importantgrowth potential. A policy ofliberalizing entry and facilitating exit has already played an important role in thisregard. Removal of constraints that underpin productivity differences amongexisting firms, including those between the state and non-state sectors willhave to be tackled next.
估计方法
(1)Benchmark measurement: 增长核算
(2)稳健性检验:Olley andPakes (1996)和Ackerberg etal. (2006)
Data
(1) 处理的三大问题
Industryconcordances, deflators for all nominal variables, programs to match firms overtime, and to construct a real capital stock series. 此外,对劳动份额的处理。We followHsieh and Klenow (2009) and inflate wage payments by a constant factor for allfirms to work with a wage share consistent with the national average (55%).
(2)数据跟踪
One-sixth ofthe Chinese firms in our sample have at least one ID change.
We use uniquenumerical IDs to link firms over time. We have aimed to track firms as theirboundaries or ownership structure changes, using information on the firm’sname, industry, address, etc., to link them.
(3)数据信息
There are148685 firms in 1998 and 313048 in 2007. On average, the annual rate ofattrition in our sample is slightly less than 14%. Out of our original samplein 1998 of 148685 firms, 33045 firms or just under a quarter, survive through2007. Exit was more than offset by entry, which averaged nearly 20% per annum.
4 results
分析框架
Aggregate TFPgrowth for the manufacturing sector can occur in three important ways:
(1)through firm-level productivity growth.
(2) exit of below-average productivityfirms or entry of above-average firms.
(3) the reallocation of productionfactors from less to more productive continuing firms.
4.2 Firm-levelproductivity growth
4.3 Enteringand exiting firms
It is alsoimportant (1) where in the distribution the new firms are entering and wherethe old ones are disappearing, and (2) how productivity growth evolves over afirm’s life cycle.
4.4 Resourcereallocation between active firms
We do this byapplying the methodology that Bartelsman andDhrymes (1998) used for U.S.plants to the Chinese firms in our data set.
4.5 From themicro to the macro level
4.5.1 Growthdecomposition
4.5.2 Stateversus non-state
4.5.3 Sectoralheterogeneity
最后一部分只列框架,公式自己到文章中看。