Types of Technical progress and its implications in Solow model

Notes based on the lecture by Prof. Alwyn Young discuss types of technical processes, Uzawa's theorem and its implications.

Click here to read my notes.

In the notes (above), we discussed Uzawa's theorem and the collorary. Also, the longer run data suggested that indeed the income shares are constant, suggesting the growth to be Harrod neutral. However, when we look at the recent data on global labour share, it is evident that this metric is actually falling. Looking at labour share for the USA, Japan, China and Germany, we observe a similar decline.

Even looking at the investment shares, we observe a significant change. Looking closely at the aggregate income shares and shares in GDP of IPP (like patents, R&D, software etc), there's a huge rise in these, especially in the post-war period.

In the earlier system of NA, IPP (the black line in the above figure) were counted as the intermediate costs. Hence, they were deducted from the income or the output. Sometimes, after the 2000s, the US (with recommendations from the UN) started accounting IPP as investment goods rather than intermediate inputs. So, this shift in accounting measures had an impact on the labour share calculations. As the figure below shows, the blue line is the labour share calculations based on the newer accounting practice (with IPP as investment goods) while the orange line (with IPP as intermediate goods) is under the older definitions. 

So, when we count IPP as an intermediate expense and not as an investment, we exclude it from all the measures of accounting, that is, expenditure, value-added and income. Consequently, these all reduce and get smaller than the case of IPP were counted as an investment good. 

Now, if we consider IPP as an investment good then it may happen that capital share is rising due to increments in IP, however, if we consider it as an intermediate good, the capital share may remain constant. This is potentially the reason that the older accounting technique showed capital shares as constant, despite the observation of increasing IPP.

Even looking at the individual countries we do see IPP rising. The capital income share (with new accounting technique), therefore also witnesses increment. Subsequently, labour income shares are not constant and actually are declining. However, it should be noted here that the companies compensation to the employees through stock options are not counted by the NA. Thus labour income shares do not take this source of income into account. If we add this metric to the labour income, then this income would rise, thereby raising the labour income shares. However, the strength of this increment in labour income shares, as a result, would depend on the magnitude of stock options given as compensation to the employees. Practically in many countries, the stock options as compensation are not as drastic and so the upward shift in the labour income share's would not be very large. Therefore, even after correcting for stock options as income of employees, the labour income shares need not remain constant but compared to the steep declines, it might get shallower. 

Moreover, there is a global trend of declining capital income taxation while labour taxation has remained more or less at the same rate. This might have implications in the way firms compensate their employees. For example, in the US consultancies, it has been brought that they pay large compensation in form of so-called "delay payments", which are accounted in NA as capital expenditures and so taxed at very low rates. This paper basically points towards the role such aspects may have in the calculations of income shares in NA. 

Looking at the trends of income in the US economy (figure above), it is clear that there indeed has been a significant rise in the income levels for the 90th percentile but barely for those in the 10th percentile. Differentiating income's between men and women show the former lagging behind the latter. Thus, there is a divergence in incomes as we compare the incomes in 1979 and 2018. Technical progress, therefore, is not "raising all boats" but "raising wealthiest boats" and creating inequalities. 

If we segregate the income trends by educational levels, we find that greater education does have an advantage. However, merely comparing median income's is not enough, as within college degree holders data shows enough dispersion. Furthermore, such representation of data shows that while college degree holders did have an improvement in income levels, individuals with high school diplomas or less suffered income contraction. Its standard explanation comes from skill-biased technical change - that is, individuals with skills get premium through the technical progress.

As a result of it, what's happening in the market is the individuals involved in the lower skill level are facing intense competition with the sophisticated technologies or machines and are getting replaced by them. In this way, such individuals getting hit to their income levels. However, the individual's who inhabit a layer above them, that is, who takes the output of these individuals as its input, is getting an efficient way of extracting input. To better understand this, take the example of X-ray technicians. It is said that soon machines would be able to read X-rays. Thus, these technicians would face competition with such machines. However, the doctors, who earlier used to rely on the efficiency of technicians would then have a more efficient machine to complement their output. In this way, individuals on the upper layer of skill-set get complimented in their productivities but those in the lower strata get hit negatively. 

Continuing the discussion on the above figures, it is interesting to witness dispersion amongst college degree holder's as well. What it says basically is that even if there are many degree holders, the income advancements would not be the same. Only, the best show significant advancement in incomes. It could again be argued on the basis of evolving technologies. For example, telecommunication technology has allowed easier broadcasts which we can sit on our sofa's and watch. So, with such ease, we would like to see the tennis match of top 1,2 or 3 players rather than player no. 10. This was not the case in earlier times when let alone watching a match was a matter of pride. Thus, only the best in the field get premiums. In this way, all these things lead to rising global inequalities. 

All these points that the Solow model, Harrod neutral progress and its basis of US growing on BGP with g(y) = g(k) are slightly misleading. However, note that it doesn't imply that Uzawa's theorem is not theoretically wrong but what it says is that there could be different types of technical progress being experienced by the economy (skill-biased, skill-enhancing, skill capital complementary etc.), that on the net at longer time scales yield constant growth paths. So, the actual pattern of what's happening doesn't look like the model of constant capital and labour income shares and all workers benefitting them from the improvements in productivity.