Download full paper

Please cite the paper as:
“Umit Senesen Gulay Gunluk-Senesen Zeynep Yilmaz, (2013), JOBLESS GROWTH WITH IMPORTED INPUTS:THE TURKISH CASE IN THE NEOLIBERAL ERA, World Economics Association (WEA) Conferences, No. 4 2013, Neoliberalism in Turkey: A Balance Sheet of Three Decades, 28th October to 16th December 2013”



The Turkish economy in the 2000’s is characterized by positive growth rates and persistent high rates of unemployment.  This paper attempts to gain insight into the concurrence of these two phenomena in the neoliberal era by taking the role of intermediate imports into account, an aspect which received relatively less attention in the jobless growth literature. Considering the upward trend of intermediate input penetration in the Turkish production sectors, we hypothesize that the same GDP growth rates are feasible with less employment generation.   Several hypothetical scenarios (e.g. substitution of 1 %, 5 % of domestic intermediate input use of all the sectors, of major exporting sectors, of highly import dependent sectors, etc.) are calibrated with the input-output data of TURKSTAT  for 2002 and estimated employment data for 31 sectors to see the extent of (possibly negative) effects on employment. The methodology provides decomposition of employment generation of policy sectors with respect to origin sectors along with decomposition of import dependency (of destination sectors).  Preliminary findings show that the shift from domestic intermediate goods to imported ones in food products, wood products, tobacco, wearing apparel, textiles and agriculture generate the highest job losses in the economy. These industries have sizeable negative effects on employment in agriculture, textiles, leather products, wood products, utilities and metals.  Hence growth performance in the neoliberal era in Turkey was not matched with employment creation the sectoral breakdown of which is identified in this paper.

5 responses

  • Cagacan Deger says:

    The article debates the jobless growth phenomenon and investigates how shifting from domestic intermediate inputs to imported intermediate inputs contributes to this phenomenon in Turkey. The investigation is conducted through an input-output analysis. One would expect a shift from domestic inputs to imported inputs to have adverse employment effects; for as the sum of intermediate and final demand falls, economy will shrink. Even though the article does not provide detailed results on production (Is output being held fixed, while intermediate input is reallocated from domestic to imported?), such expectations seem to be confirmed.

    Even though the article is a decent contribution, this is not the single explanation on jobless growth. The authors freely admit this. I believe for the Turkish case that two issues need to be addressed as well. Firstly; Turkey’s growth is frequently related to availability of foreign financial resources. Turkish economy grows as long as there are channels for foreign borrowing. Such an issue can hardly be addressed through an I-O framework; still, it needs to be considered in relation to jobless growth. Secondly, anecdotal evidence implies that a more “efficient” employment of the existing workers (i.e. keep a portion of pre-crisis labor force, have them work more) seems an equally important explanation. I think this article did not resolve the issue of jobless growth, but added one more possible explanation to a stack of explanations. I hope that a perspective that integrates and compares existing explanations will arise.

  • Osman Aydogus says:

    The paper provides valuable empirical insight into the jobless gowth phenomena observed in the neo-liberal era in Turkish economy. The I/O method used in the analysis not only gives a possibble eplanation for the jobless growth in Turkish economy after 2000 in general, but also identifies job creation (losses) under ‘the same growth rates but smaller direct domestic input coefficients’ (higher direct import coefficients) on sectoral level. The methodology takes into consideraton the industrial interdependeny among the sectors and hence allows to distinquish between direct and indirect effects of a rise in direct import coefficients (a shift from domestic to imported inputs) on individual sectors’ output and employment levels.
    Analysis depends on TURKSTAT’s 2002 Turkey IOT, the latest availible official IOT. It is very likely that during the last decade some miner changes in the total direct input coefficients and significant changes in the composition of domesticly produced and imported inputs have occurred. One can expect higher direct import coefficients, especially in sectors where import dependency on intermediaries deeepened, such as electricity generation, oil refining, basic chemicals, recycling, wood products, food, tobacco, wearing apparel, textiles, etc. However, the likely changes in the composition of direct input coefficients during the last decade do not invalidate the main findings of the anlysis: “Empirical findings suggest that shifting from domestic to imported intermediate inputs weakens employment generation potentials.”
    Among others, one conclusion derived form the empirical finding is crucial: “… targeting on policy or origin sectors would be an inadequate approach for solution. Instead it is necessary to take into account employer sectors associated with policy sectors. Thus sectoral composition of public investments and of private investment incentive schemes is a crucial issue to be integrated into policy design for employment creation. However, this integrated approach for policy and employer sectors has long been overlooked by the neoliberal policy design, Turkey being no exception.”
    The study is a very valuable empirical contribution, providing insight into the possible effects of intermediary imports penetration on sectoral output and employment losses during the last decade.

  • Osman Aydogus says:

    Dear Cagacan,
    Although not clearly stated and not detailed, the I/O methodology used is obvious. Equilibrium output employment levels are calculated for the same final demand vector (corresponding to a definite, although not clearly stated in the article, GDP growth rate) for two sets of direct input coefficients: 1) the actual 2002 direct domestic and import input coefficients, and 2) higher direct import coefficeints. In other words, by the logic of static-open I/O methodolgy, causation runs from changes in the final demands towards changes in output and hence employment levels.
    You state that “Turkey’s growth is frequently related to availability of foreign financial resources”. You are quite right. The positive correlation between growth and foreign financial flows is another well established and generally accepted fact in Turkish growth literature. We also know that the foreign resources are used to finance imports. Then we can suggest a link between import penetration and availibility of foreign financial resources, although it can not be adressed formally within the static-open I/O models.
    I agree that “more “efficient” employment of the existing workers seems an equally important explanation”. However this issue can be addressed in the model simply by reducing the direct employment coefficients.

  • Cagacan Deger says:

    Dear Prof Aydogus,

    Thank you for the clarifications on the article. I had difficulty in seeing exactly how growth was introduced into the model. Having clarified that, I am now concerned about the implications of such a growth formulation.

    My understanding at this point is that given a final demand vector (which implicitly accounts for growth, being calculated through GDP growth rate) the article analyses the impact of shifting from domestic to imported inputs. And this is performed within a static-open I-O model. It is most logical to have falling employment in such a framework. For given L/X rates, falling demand for domestically produced intermediate demand would lead to falling employment. This implicitly includes falling sectoral output.

    However, are we seeing the full effects of the introduced shock? It is an open model; I take this to mean that final demand is exogenous. What if final demand (consumption, for example) is endogenized and thus the model is closed? Then we could see the induced effects of the shock (shock => production changes => income changes => consumption changes => a second round of final demand shock …). It is possible to obtain falling output rather than growth in a closed model. The initially assumed exogenous growth could easily be reversed. Then we end up with a model which shows that shifting to imported intermediate goods from domestic intermediate goods simply reverses growth. Then the model can not account for jobless growth.

    So the question is: Is there anything in the model that excludes such possibility of reversed growth? Or am I missing yet an other important point?

  • Ebru Voyvoda says:

    The manuscript takes the issue of “jobless growth” and the relationship of the issue with increased reliance on imported intermediates. The topic chosen is quite interesting since the phenomenon of jobless growth on the one hand, and the increased reliance of the production structure of the economy on imported intermediates, on the other are highly controversial and intensely discussed topics within the Turkish context. The authors in the manuscript cite the important contributions.
    The manuscript tries to add to this literature by making use of the Input/Output (I/O) tables and the I/O analysis. To this end, the authors disaggregate the I/O matrix (A) into its domestic (Ad) and imported (Am) components and reach to Equation (9), which links the (gross) output produced to the domestic demand (yd) through matrix R, the Leontief Inverse with the domestic coefficients. Then, the (fixed) labor coefficients (L) are used to transform the output (x) to different categories of labor demand.
    The method in fact, at this point calls for the step, where changes in x (Δx) to changes in the I/O coefficients (the A = Ad+Am matrix) and the changes in the components of final (domestic) demand. Then, one would get the complete picture of the effects of changes in the production structure of the economy and the changes in the components of final demand on the output; and through labor coefficients on labor demand. Of course, such a complete analysis calls for two I/O matrices for two separate years: It would be a complete analysis if one had not only the 2002 I/O Table but also a recent one.

    The issue here is then, whether the partial analysis here (relying only on the 2002 I/O Table) can give some insight to understanding the relationship between jobless growth and the increased reliance on imported intermediates. My answer would be “very vague” since one has to rely on very strong set of assumptions. These assumptions are especially strong for a period where the overall structure of the production and sectoral distribution of both production and employment are changing:
    • How relevant it would be for the Turkish case to assume that the I/O structure of the economy has not changed since 2002?
    • Given the changing composition of exports, for instance, one can assume that there has been significant shifts in the sectoral production and employment even if the I/O structure stayed constant. So, any shift towards intermediate imports could have been compensated by the shifts in demand to imply higher employment. Unfortunately, one cannot decompose and see the effects of each component.
    • The analysis and the results cannot be specific to Turkish economy as Equations (9-10) imply: for any economy any shift from imported to domestic intermediates would create direct and (through sectoral linkages indirect) stimulus on domestic production, whereas the reverse would cause a movement in the opposite direction. The relationship is clear-cut and would be valid for any I/O Table. So, one can expect nothing but only a negative effect on (domestic) output and employment by the scenarios carried out in the manuscript.
    • I think, what the authors are aiming in this manuscript is important, but by relying on I/O Table of one year, one cannot have the direct implication of intermediate imports and jobless growth phenomena. The analysis here only takes the “intermediate import” part, but I think leaves out the “growth” and “sectoral and final demand distribution of growth” out.

    I thank the authors for the contribution for I think the method here can serve well to see the overall linkage effects of the economy and changes in production and demand structure on employment.