We have launched E-mail Alert service,subscribers can receive the latest catalogues free of charge

 
 

Problems with Current TFP Measurement and Suggestions on Improvement

2019-04-16

By He Jianwu

Research Report Vol.21 No.2, 2019

As Chinese economy has moved from the stage of high-speed growth to one of high-quality development, improving the quality and efficiency of growth is a pressing issue. The report delivered at the 19th CPC National Congress made it clear that we will “work hard for better quality, higher efficiency, and more robust drivers of economic growth through reform and raise total factor productivity.” Total factor productivity, or TFP, an indicator that gauges growth efficiency, has become one of the main indicators of high-quality development that is watched closely. Accurately measuring TFP not only concerns policy formulation for high-quality development, but also affects the evaluation of policy effect. This paper tries to analyze possible problems and challenges of the current TFP measurement from various perspectives, including traditional accounting method, service sector, digital economy and new development concepts, and aims to inform decision-making for better supporting high-quality development.

I. To Advance High-quality Development, We must Measure TFP Changes and its Contribution to Economic Growth more Accurately

On the supply side and in a long-term view, economic growth is driven by factor input and raised productivity. The remainder of economic growth after deducting the increase in all factor input (labor, material capital, etc.) is called TFP growth, which is usually taken as an important indicator of efficiency and technological improvement. By comparing the economic growth contribution from factor input growth and TFP growth, we can assess whether the economic growth is quantitative expansion (which mainly relies on factor input) or driven by quality and performance improvement (which mainly relies on efficiency and technological improvement).

China’s high-speed growth in the past 40 years was mainly quantitative expansion that relied on the massive input of cheap factors, whereas efficiency and technological improvement contributed relatively little. Empirical studies show that capital input contributed nearly 60% to economic growth in the past nearly 40 years, labor force contributed nearly 10%, and TFP contributed only 30% or so as opposed to the usual 40%-50% or more in developed countries. In promoting high-quality development, one of the key tasks is pushing economic growth to shift from quantitative expansion to quality and performance improvement, or, in essence, increasing TFP’s contribution to economic growth. Therefore, accurately measuring TFP changes and its contribution to economic growth is of significant import for advancing high-quality development.

However, as Robert Solow (1987)[] said, we saw computer everywhere, but we didn’t see it in productivity statistics. The result of TFP measurement usually differs from the productivity changes we observe in real life. There are many reasons for this, such as disparity in perception (partial VS whole, innovation VS TFP), the measurement not being objective enough, subjective reasons on the measurer’s part, limitation of measuring method, defects in statistical system and method, and flawed TFP measuring method.

II. Traditional Accounting Methods and Data Undermine the Accuracy of TFP Measurement

Historical comparison shows that TFP measurements conducted by different scholars usually differed wildly, which is to a large extent caused by traditional accounting methods and data. Specifically, this involves three aspects.

1. TFP has obvious “pro-cyclical defect” when measured with all capital stock

Most growth-based TFP measuring methods now use the total capital stock rather than effective capital (capital that’s actually invested in production). At times of economic slowdown, capital utilization rate is generally low; there is a lot of idle capital and what’s actually invested in production is less than total capital stock. At times of economic boom, capital utilization rate is usually high; a lot of machines and equipment are in overloaded operation and what’s invested in production may be more than total capital stock. Therefore, if capital utilization rate is not considered, TFP may be underestimated during economic ebb and overestimated during economic boom, which is obviously pro-cyclical. This not only affects the accurate measurement of TFP, but also undermines the indicator’s guiding effect.

2. TFP underestimates the speed of technological progress when measured in disregard of investment-specific technological changes

Technological progress is not only reflected in knowledge and experience accumulation, but also in materialized capital, but traditional growth theories and accounting methods usually focus on the disembodied and neutral technological progress, so they cannot reflect the efficiency improvement resulting from investment-specific technological progress[] (machinery and equipment). Whether late-starting countries can differentiate it from neutral technological progress directly concerns whether their policy suggestions are scientific. There have long been discussions about this topic[] and mature solutions have been created. We can modify the equation of capital accumulation to introduce investment-specific technological progress, meaning we measure the quality of capital with efficiency unit rather than the traditional physical unit. In other words, assume that new capital is more efficient than old capital, so we endow them with different efficiency factors. Data about capital quality is lacking in China[], so investment-specific technological progress is rarely measured, and a consensus is hard to reach.

3. Discrepancy in the measurement of factor input leads to disparity in TFP measurement

The accuracy of factor input measurement directly concerns TFP accuracy, and different results of the former lead to huge disparity in TFP results. It’s hard to unify the measurement of factor input due to a spate of complicated reasons, mainly reflected in the following aspects. First, the amount of labor force input. Most measurements now use the total employment, but laborers’ average working hours vary from region to region, from period to period. Compared with foreign workers, it’s more common for Chinese workers to work overtime. As a result, if total employment is taken as labor force input, the efficiency in China may be overestimated as the extra output is attributed to overtime work rather than higher efficiency. Second, the amount of capital input. Apart from the issue of valid capital mentioned earlier, a main problem with capital input measurement lies in the estimation of capital stock, especially the definition of depreciation rate. Based on existing estimation of capital stock in China, the depreciation rate varies vastly from 5% to 14%, and its diversity directly affects the estimation of capital stock and its growth rate, consequently the TFP measurement. Third, the quality of input factors. There is quality difference within each factor, such as laborers of different educational background or skills. An indiscriminate summation would affect the accuracy of TFP measurement. Besides, how to reflect the betterment of product quality in output measurement is a chronic statistical issue that is yet to be well solved.

...

If you need the full text, please leave a message on the website.