Productivity in Ireland 2019

Get the data: PxStat PIA10

The chart above depicts Gross Output in the Construction Sector broken down by its contribution from KLEMS inputs. It is compared with its GVA counterpart below:

Get the data: PxStat PIA09

The main difference between the KLEMS and GVA models of productivity is the inclusion of intermediate inputs in a more detailed series that estimates MFP based on Gross Output rather than on GVA in any given industry. The two charts above compare the KLEMS and GVA models of productivity for the Construction sector. The KLEMS model offers a more detailed analysis in terms of what is driving growth and productivity in the industry. Along with Labour and Capital, additions to Energy, Materials and Services can also explain economic developments in this sector in the KLEMS model. This level of insight is not available from the GVA model of the Construction sector which reports a relatively large MFP.

Value-added MFP measures the unobservable drivers of productivity not related to the primary factor inputs of Labour and Capital and is sometimes known as the Solow residual. The KLEMS model minimises the MFP estimate where MFP = Output – Capital – Labour – Intermediate Inputs while the GVA model excludes these intermediate inputs and, in many instances, results in larger MFP. In theory, the inclusion of the additional intermediate inputs in the KLEMS model should yield a more accurate MFP estimate as it minimises the Solow residual by including more explanatory variables.

In practice, the inclusion of intermediate inputs sometimes makes the KLEMS results more susceptible to data quality issues and for this reason the KLEMS MFP estimates are considered supplementary and experimental. The Gross Output series in the KLEMS model is considerably more volatile than the GVA model by nature; the KLEMS series uses a gross measure for output while GVA is in practice a net measure.

In terms of the data sources, both KLEMS and GVA methodologies estimate Capital Services using the CSO's Capital Stock of Fixed Assets and Labour estimates are sourced from the Labour Force Survey (LFS). The addition of intermediate consumption estimates from the Supply and Use Tables (SUT) was necessary for estimating MFP on a Gross Output basis in the KLEMS model. Using the SUTs in the KLEMS analysis may have also contributed to volatility and certainly highlighted data gaps where annual SUTs don’t exist for some earlier years.  This necessitated additional estimation and adjustment in the creation of a complete time-series. For these reasons, GVA estimates are considered to be more robust and deliver more consistent estimates for productivity analysis and the KLEMS based productivity estimates are considered experimental for Ireland.

Data Sources

In relation to data sources, as indicated above Capital and Labour inputs are sourced from the CSO’s Capital Stock publication and LFS which are also used throughout the publication in all productivity presentations. The intermediate inputs are obtained from the SUTs and used together with Capital and Labour inputs to estimate productivity on a Gross Output basis. The SUTs provide a detailed picture of the transactions in goods and services by industries and consumers across the Irish economy in a single year. They also highlight the inter-industry flows that lie behind the National Accounts main aggregates. The Supply table contains estimates of the supply of goods and services (products) by domestic industries as well as imports of goods and services while the use table contains estimates of the use of products by domestic industry and by the final demand sectors[i]. The SUTs are broken down using the NACE Rev. 2.0 classification for economic activity with a 21 sector presentation of the economy, results can be found on PxStat.

[i] (Final Demand comprises consumption by households, government, non-profit organisations serving households (NPISH), gross fixed capital formation (GFCF) and exports)

KLEMS - The Challenges

The production of a single consistent presentation of the input data for Energy, Materials and Services from the SUTs for the entire nineteen years was therefore challenging due to gaps in the data series for some years.  There were also major changes in the National Accounts Methodology and Classifications. The years 2000-2019 witnessed the reclassification of economic activity from NACE Rev. 1.1 to NACE Rev. 2.0 and the change from the European System of Accounts (ESA) 1995 to ESA 2010.  Accordingly the production of the estimates presented here entailed several steps to ensure data quality and consistency[1] between the KLEMS and GVA frameworks. The series is nevertheless considered experimental.

[1] The Domar methodology was used to validate the KLEMS results in the publication. See Appendix for more information.

Get the data: PxStat PIA10

The above chart shows Gross Output growth in the Electricity, Gas & Steam sector broken out by the KLEMS inputs. Unsurprisingly, the largest contribution to Gross Output over the entire period came from Energy. Looking at 2019, the contribution of Energy to Gross Output growth was 2%. However, at times the trend in the Energy input seems to be not entirely in line with the economic cycle, particularly in 2008. Recent years have seen increased contributions from Materials and Services and declines in the contribution of Energy. As it is important to consider the factors that affect the costs of intermediate inputs, i.e. producing electricity, further analysis of the Energy market are presented below:

Get the data:  PxStat PIA10

The chart above presents growth in the Energy input of the Electricity, Gas & Steam sector against oil prices, using an annual dataset from the Federal Reserve of Economic Data (FRED) on Brent Crude Oil prices. Oil or products indexed to oil prices such as gas, are key inputs in electricity generation. It was found that there was a high positive correlation of 74% between Energy inputs in Electricity, Gas & Steam and Oil Prices ($ per barrel) during the years of 2000-2019. Oil prices increased by 29% in 2008 and this may be what led to the large change in Energy inputs in 2008 in the sector. 

Oil prices have been volatile, however the declining trend in the oil price is associated with falls in the contribution of Energy to the Electricity sector. In more recent years the cost structure of Energy input is more difficult to analyse with the increased contribution of renewable energy.

Get the data: PxStat PIA10

The chart above breaks Gross Output growth down into the factor inputs in the Professional, Scientific, Admin & Support Services sector for the years 2000-2019. Gross Output declined in 2019, when compared to 2018, falling from 11% to 3% with the majority of the contributions from Labour and Capital. In general as this sector is predominantly a services-based one, swings in Gross Output are largely explained by changes in the services contribution, however this has become less pronounced in the most recent years. Gross Output growth declined from 31% in 2007 to -26% in 2011 and can be attributed to declining contributions from Materials, Services and Capital. This fall in Gross Output between 2007-2011 was likely caused by the economic recession period as high skilled industries, which have long lead in projects, are likely to witness a gradual decline in output rather than a sudden drop which was seen in industries like Construction and Accommodation and Food. Gross Output in this sector recovers after 2011 and records positive growth every year since 2012, with an average growth rate of 2.7% recorded between 2010-2019.

Get the data: PxStat PIA10

The changes in Gross Output growth between 2000 and 2019 in the Financial and Insurance sector can largely be explained by changes in Services Input, MFP and Capital. Gross Output growth in the sector for 2019 was 6% and was mostly due to the contribution of 5% from Capital. On the other hand, the decline in Gross Output from 6% in 2007 to -4% in 2008 was mainly due to the fall in the Capital contribution from 6% to 3%, as well as a dip in the Services input. Gross Output was slow to recover in the Financial and Insurance sector as this period was characterised by what was termed a Balance Sheet recession where a period of asset inflation was followed by a crash in these same asset values. The recovery period saw drastic changes in financial regulation as the Central Bank moved towards adopting and implementing macro-prudential policies and this explains, to some extent at least, the slow and steady recovery. Since 2017, growth in Gross Output has averaged 5.6% per annum, due to strong contributions from Capital and Services.

Get the data:  PxStat PIA10

Growth in Gross Output and productivity in the Accomodation and Food sectior are largely associated with changes in Labour and Materials consumed in the production process. Since 2017, growth in Services and Labour have averaged approximately 2-3% per annum. The industry experienced a sharp decline in Gross Output after the economic crisis in 2008. Gross Output growth plummeted from 16% in 2008 to -15% in 2009 at the height of the recession and remained negative until 2014. The period also saw lower levels of overall activity and discretionary consumption as many household budgets were impacted by the events of 2008. This resulted in lower levels of demand for Accommodation and Food services. Since 2016, growth in Gross Output has remained positive largely explained by increases in Labour and Services inputs. As household and corporate budgets have increased in recent years, this has led to an increase in demand for Accommodation and Food services.

Get the data: PxStat PIA10

The chart above decomposes Gross Output growth in the ICT sector into its factor inputs and intermediate inputs. The key input in the production process is Services over the entire period, however on account of large imports in 2017 and 2019, Capital played the dominant role. Gross Output growth in the ICT sector for 2019 was -1.8%, down from 25.7% in 2018. Despite the overall fluctuations for the ICT sector in the Gross Output, GVA growth has remained consistently strong over much of the entire period, with a result of almost 16% in 2019. 

Get the data: PxStat PIA10

The shares in Gross Output of Energy, Materials, Services, Capital and Labour for the Accommodation and Food sector over the period 2000-2019 are shown in the Figure 9.9. The most significant shares are Labour, Services and Materials, however in recent years the role of labour is becoming more significant in the sector. The Labour share has remained consistently around 50% since 2012. There appears to be some substitution between Materials and Labour in the sector, as the rise in the Labour share has coincided with a relative decrease in the Materials share, falling from 29% in 2010 to just 9% in 2019. For the period since 2008, a Services share of over 20% was reported.

Get the data: PxStat PIA10

The shares in Gross Output of Energy, Materials, Services, Capital and Labour for the Manufacturing sector over the period 2000-2019 are shown in Figure 9.10. Capital and Services have the most significant share, with the Capital share increasing sharply since 2015. The Capital share in 2019 was 53%, while the Services share was 35%. The impact of the additions to the capital stocks in 2015 in the Capital share resulted in reductions in all other Gross Output shares. The labour share has stayed around 7% since then and this is an indication of the capital-intensive nature of this multinational-dominated sector.