Exchange-rate pass-through in Turkey with a supply and use model

ABSTRACT

Using a multisectoral model and data from the supply and use tables, this article estimates the exchange-rate pass-through to the domestic price level of the Turkish economy. The results indicate that a unit change in the exchange rate leads to 0.61 units of change in the overall price level. This effect comes out mostly through the industry, and the primary sector; while in terms of products, the energy sector is the major actor for the changes in the domestic price level.

KEYWORDS:

• CPI
• exchange-rate pass-through
• inflation
• supply and use tables
• Turkey

JEL CLASSIFICATION:

• C67
• D57
• E61
• F41

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ The ERPT refers to changes in prices due to changes in the relative value of the domestic currency vis-à-vis other currencies.

² For details of background information on exchange rate and other determinants of Turkey’s economy health, see Akgunduz and Fendoglu (2019).

³ It should be noted that in Aydoğuş et al. (2018) framework the estimates “should be considered to be upper limits; actual changes in the prices are likely to be lower” (p. 5).

⁴ For the described products and their correspondence to Classification of Products by Activity (CPA), see https://data.tuik.gov.tr.

⁵ Matrices (and vectors) are delineated in boldface letters. The $1\times n$ vector is denoted by $\mathbf{x}\equiv \left[x_j\right]$ , and the diagonal matrix formed from the elements of $\mathbf{x}$ is denoted by $\hat{\mathrm{x}}$. Finally, $\mathbf{i}$ denotes the summation vector, i.e. $\mathbf{i}\equiv \left[1,1,\dots ,1\right]$ and ${\mathbf{i}}_j$ the $j-th$ unit vector.

⁶ The part of the supply table that describes domestic production is considered as the empirical counterpart of the output matrix in our modelling, while the part of the use table that describes intermediate consumption (domestic and imported) is considered as the empirical counterpart of the input matrix (see, e.g. Bidard and Erreygers 1998, 434–436).

⁷ For a detailed discussion see, e.g. United Nations (1999).

⁸ As Aydoğuş et al. (2018) note, $\mathrm{d}{\mathbf{p}}^{\mathrm{f}}={\mathbf{p}}^{\ast }\mathrm{d}\mathrm{e}+\mathrm{e}\mathrm{d}{\mathbf{p}}^{\mathbf{\ast }}$, and given that $\mathrm{d}{\mathbf{p}}^{\mathbf{\ast }}=0$ , $\mathrm{d}\mathrm{e}=\mathrm{d}{\mathbf{p}}^{\mathbf{f}}/{\mathbf{p}}^{\mathbf{\ast }}$.

⁹ For two different interpretations of the sums of the column elements of $\mathbf{M}\left(\mathbf{I}+\hat{\mathbf{r}}\right)]{\left[\mathbf{B}-\left(\mathbf{A}+\mathbf{T}\right)\left(\mathbf{I}+\hat{\mathbf{r}}\hspace{0.17em}\right)\right]}^{-1}$, see Hummels, Ishii, and Yi (2001) and Reis and Rua (2009).

¹⁰ Considering no impacts of the central bank adjustments in its interest rate; no substitution between imported intermediate inputs and domestically produced intermediate inputs; the cost increases are proportional to prices, and that the price increase in consumer goods and services are reflected proportionally in wages, our estimates should be considered to be “upper limits” (see, also, footnote 3).

¹¹ Note that the definition of several sectors has changed between 2002 and 2012, thus the differences focus only on the sectors the definition of which has practically remained the same.