House prices and labour productivity growth: Evidence from OECD countries

Abstract

This study examines the association between real house prices and labour productivity growth in 24 OECD countries over the period 1972–2019. By applying the panel fixed effects and Pooled Mean Group-Autoregressive Distributed Lag (PMG-ARDL) estimators, the results show that labour productivity growth is negatively and significantly associated with real house prices. This finding provides empirical support for the ongoing discussion on the adverse impact of excessive housing market activities on productivity.

Keywords:

• House prices
• Labour productivity
• OECD
• Panel data

1. Introduction

In recent years, a number of conceptual and theoretical studies have argued that housing market outcomes (e.g., appreciation in house prices and housing investment) can have adverse impacts on labour productivity growth (for a comprehensive review of literature, see Dodson et al. 2017; Maclennan et al. 2015; Maclennan et al. 2018; Maclennan et al. 2021; Pawson et al., 2021). It is argued that house price appreciation and growth in property investment have led to impaired labour productivity (and ultimately slower economic growth) through, at least, three channels: (1) “The productivity impairment due to the housing system’s tendency to distance homes affordable to low and middle income earners from employment growth hubs” (Pawson et al., 2021, p.6). (2) “The opportunity cost arising from the channelling of debt-fuelled investment into housing stock – an asset essentially unproductive in terms of employment generation” (Pawson et al., 2021; p.6). In other words, it is argued that since real estate services and construction have smaller than average sectoral (e.g., manufacturing) productivity¹, then increased housing investment may lower productivity² (Pawson et al., 2021; RBA 2019). (3) “The high proportion of income spent on housing, especially for renters, thus precluding expenditure on other consumption items, and thereby reducing overall demand in the economy” (Pawson et al., 2021; p.6).

However, on the other hand, there is a thread of literature which argues housing market activities (e.g., housing design, building, financing, transaction, maintenance and renovation) significantly contribute to economic activities and productivity. In addition, house price appreciations play crucial roles in the development of small businesses (through housing collateral and housing wealth effects) which are sources of innovation and productivity growth. Housing wealth and housing collateral effects are also key drivers of household consumption Gholipour and Tajaddini (2017) which lead to higher levels of firms’ production (for a review, see Gholipour 2020; Maclennan et al. 2018; Pawson et al., 2021; Maclennan et al. 2021). In other words, although the productivity is low in real estate services and construction sectors, it is important to recognise the long-term positive impact of these sectors in building productive assets which rematch households and residential properties in ways that increase productivity (Pawson et al., 2021).

While the existing studies provide valuable theoretical insights on the link between housing and productivity, there is a lack of empirical macro-level research investigating the relationship between real house prices and labour productivity growth across countries and over time. The gap in the literature is addressed by applying the panel fixed effects and Pooled Mean-group/Autoregressive Distributed Lag (PMG-ARDL) estimators on a dataset of 24 OECD economies for the period from 1972 to 2019. Our study is also motivated by the call for further study on the causal relationship between productivity and housing market indicators (Maclennan et al. 2021).

Our estimation results show that labour productivity growth is lower in countries with higher real house prices. This finding indicates that housing market activities should be considered as a significant determinant of productivity growth in developed economies by economic policymakers (Maclennan et al. 2021), along with other key determinants of labour productivity (e.g., capital deepening and education).

The focus of this study is on labour productivity growth because it is a key element of economic performance and a crucial driver of variations in living standards (OECD 2022a). According to Paul Krugman, “… a country’s ability to improve its standard of living over time depends almost entirely on its ability to raise output per worker’’ in the long run (Krugman 1994). However, productivity growth has been sluggish across advanced and emerging economies in recent decades (see Figure 1, adapted from Dieppe 2020), and therefore the issue has been a central priority for the global development agenda (Colford 2016).

Figure 1. Global, advanced economies, and emerging market and developing economies (EMDEs) productivity growth.

Figure 1 Productivity is defined as output per worker in US dollars (at 2010 prices and exchange rates). The sample used in drawing Figure 1 comprises 29 advanced economies (AEs), and 74 emerging market and developing economies (EMDEs) including 11 low-income countries (LICs), as of 2019 World Bank classifications, 52 commodity exporters and 22 commodity importers. Aggregate growth rates are GDP-weighted at constant 2010 prices and exchange rates. Shaded regions indicate global recessions and slowdowns (1982, 1991, 1998, 2001, 2009 and 2012), as defined in Kose and Terrones (2015) and Kose et al. (2020). Available at Ch. 1: Global Productivity Trends: https://www.worldbank.org/en/research/publication/global-productivity

Source: Conference Board; Penn World Table; World Bank, World Development Indicators.

The article proceeds as follows: Section 2 reviews the related studies; Section 3 explains the data, variables and the estimation method; Section 4 reports the findings; and the final section concludes the paper.

2. The link between housing market outcomes and productivity

The literature has provided at least three channels wherein changes in housing market activities affect labour productivity in the economy. The comprehensive review of existing works in this area can be found in Dodson et al. (2017), Maclennan et al. (2015), Maclennan et al. (2018), Maclennan et al. (2021) and Pawson et al. (2021). Here we briefly highlight the key messages of each channel.

2.1. Capital allocation

Based on a review of existing studies in both advanced countries and emerging economies, Maclennan et al. (2021, p.39) argue that “Rising housing prices impact resource allocation through both the collateral channel (households using their increased housing wealth to take additional borrowing to fund non-housing investments) and the investment channel or crowding-out channel (where there is diversion of investment flows to housing from other, more productive, innovative or entrepreneurial, destinations such as business start-ups or rapidly growing businesses or economic sectors).” The first channel may promote productivity of the whole economy by enabling collateral constrained potential entrepreneurs to establish new businesses (e.g., Schmalz et al. 2017; Corradin & Popov 2015) which often invest in non-real estate capital investment and innovate products. However, the investment channel may have a harmful impact on productivity by diverting capital investment into real estate which are considered as low productive assets (Doerr 2020; Jordà et al. 2019; Miao & Wang 2014; Rong et al. 2016).

2.2. Labour mobility

House price changes affect labour mobility and labour market matching (Liu et al. 2022). It is argued that rising house prices and rents (generally housing costs) can drive employees and businesses out of the most innovative clusters (or productive locations) (Ong & Leishman 2020). In other words, high housing costs may force lower income families further away from regions with high job concentrations, that can weaken labour market matching effectiveness and subsequently lead to lower labour productivity (e.g., Maclennan et al. 2021; Leishman et al. 2021; Maclennan et al., 2019).

2.3. Formation and use of human capital

Rising housing costs take funds out of households’ investment in education and training which are the key determinants of labour productivity. In other words, high housing costs (either high mortgage repayments or rent) can impair individuals’ economic capabilities and labour productivity, in particular among poorer households.

In addition, the increases in property prices can have some productivity benefits (Maclennan et al. 2021), as “rising levels of mortgage indebtedness appear to be extending working lives…[which] will help mitigate declining rates of employment and productivity slowdown due to population ageing” (Cigdem-Bayram et al. 2017, p.1). However, some studies (e.g., Atalay et al. 2016) provide evidence that rising house prices (and growth in housing wealth) reduce labour participation in the labour market and hours of work for homeowners (especially older females and younger partnered people) which lead to a reduction in labour supply and slower labour productivity growth.

Based on the discussion above and since most arguments are in favour of adverse impact of housing market outcomes on productivity, we argue that higher levels of real house prices have led to slower labour productivity growth.

3. Data, variables and estimation method

The sample includes 24 OECD countries for the period of 1972–2019. It includes all those economies for which information on housing market outcome (real house prices) as well as labour productivity growth are accessible. The sample economies are Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Great Britain, Greece, Ireland, Israel, Italy, Japan, Luxemburg, Netherlands, New Zealand, Norway, Portugal, South Korea, Spain, Sweden, Switzerland, and the US. Beside data availability for OECD countries, this set of OECD countries play an important role in producing global outputs as they account for about 55% of the world’s GDP (constant 2015 US$) in 2021 (World Bank 2023a). In terms of geographical distribution, the sample includes countries from different regions (Asia, Europe, North America, and Middle East). Additionally, it consists of countries that have different levels of labour market competitiveness. For example, World Economic Forum (2019) ranked Switzerland, Denmark, the US and New Zealand (as well as Singapore) among the top five countries in terms of labour market competitiveness (including internal labour mobility sub-index). On the other hand, Greece, Italy and South Korea are ranked as 111, 90 and 51 out of 141 economies in 2019³.

3.1. Dependent variable

Labour productivity growth is the dependent variable in this study. The OECD (2022a)’s measure of labour productivity annual growth rate is used. OECD (2021) computes two measures for labour productivity at the total economy level: (1) Gross domestic product (GDP) at market prices per hour worked and (2) GDP per person employed. For the first one, the measure of labour input is the total number of hours actually worked by all persons engaged in production (i.e., both employees and the self-employed). For the second one, the measure of labour input is the total number of persons employed (i.e., both employees and the self-employed). OECD prefers GDP per hour worked to GDP per person employed in their analyses. Following the OECD approach, we also use GDP per hour worked in our estimations. Table 1 shows more detailed explanations of the variables and data sources. The descriptive statistics of all variables are presented in Table 2.

Table 1. Descriptions of variables and data sources.

Variable	Definition and data sources
Labor productivity	GDP at market prices per hour worked growth rate (%). Source: OECD (2022a)
Real house prices	“. the ratio of the nominal house price index to the consumers’ expenditure deflator in each country…” (2015 = 100). Source: OECD (2022b)
KOF globalization	The annual change of economic, social and political dimensions of globalization. Source: Gygli et al. (2019)
Tertiary education	The annual change in school enrollment, tertiary (% gross). Tertiary education, whether or not to an advanced research qualification, normally requires, as a minimum condition of admission, the successful completion of education at the secondary level. Source: World Bank (2023b)
Investment	Gross capital formation (annual % growth). Annual growth rate of gross capital formation based on constant local currency. Aggregates are based on constant 2015 prices, expressed in U.S. dollars. Gross capital formation (formerly gross domestic investment) consists of outlays on additions to the fixed assets of the economy plus net changes in the level of inventories. Source: OECD (2022c)
Democracy	The annual change in participatory democracy index. The participatory principle of democracy emphasizes active participation by citizens in all political processes, electoral and non-electoral. To make it a measure of participatory democracy, the index also takes the level of electoral democracy into account. Source: Coppedge et al. (2023)
Property rights	The annual change in property rights. Private property includes the right to acquire, possess, inherit, and sell private property, including land. Source: Coppedge et al. (2023)
Working age population	The annual changes in total population between the ages 15 to 64 as a percentage of the total population. Source: World Bank (2023c)

Table 2. Descriptive statistics.

Variable	Obs.	Mean	Std. dev.	Min	Max
Labor productivity	848	2.02	2.28	−7.06	20.64
Real house prices	848	79.23	33.03	24.42	169.24
KOF globalization	848	0.71	1.25	−4.39	8.40
Tertiary education	848	1.32	2.92	−50.55	25.86
Investment	848	2.67	7.59	−25.37	100.94
Democracy	848	1.12	12.16	−11.74	287.50
Property rights	848	−0.10	2.80	−38.66	24.00
Working age population	848	0.02	0.37	−1.30	1.32

3.2. Explanatory variable of interest

Our main variable of interest is real house prices. We use OECD (2022b) real house prices index (2015 = 100) as a measure of house prices in this study.

3.3. Control variables

To test the relationship between real house prices and labour productivity growth, it is essential to control for other key drivers of labour productivity growth. In the empirical estimations, we control for KOF globalization, tertiary education, investment, democracy, property rights, and working age population. It is expected that labor productivity growth is higher in economies with higher level of integration with the rest of the world through e.g., international trade and investment (e.g., Mallick 2015), higher ratio of people with tertiary education (Dieppe 2020), more capital investment (Dieppe 2020), quality of institutions including deeper democracy (Rodríguez-Pose & Ganau 2022), more respect for property rights (e.g., Carbonara et al. 2021), and a higher working-age population share (Dieppe 2020).

3.4. Estimation method

We apply a panel fixed effects model for estimations (see Eq.1(1) $${\text{LPG}}_{\mathit{\text{it}}}={\beta }_1{\text{HP}}_{i,t}+{\beta }_2{\mathrm{X}}_{\mathit{\text{it}}}+v_{\mathrm{i}}+u_{\mathit{\text{it}}}$$(1) ). Fixed effects control for any country-specific characteristics that may affect labour productivity growth (e.g., geography, culture, religion, among others). (1) $${\text{LPG}}_{\mathit{\text{it}}}={\beta }_1{\text{HP}}_{i,t}+{\beta }_2{\mathrm{X}}_{\mathit{\text{it}}}+v_{\mathrm{i}}+u_{\mathit{\text{it}}}$$(1) where LPG represents labour productivity growth; HP is real house price index; X is a vector that includes the control variables; v captures the country-specific effects⁴; i = 1,…, n denotes the country; t = 1, …, t denotes the time period; βs are coefficients; and u is an error term. We use robust standard errors in reporting t-statistics. These standard errors are adjusted for clusters at the country level.

4. Results

4.1. Main analysis

Table 3 reports the results of panel fixed effects estimations. The analysis indicates that there is a significant and negative relationship between real house prices and labour productivity growth in the sample of OECD countries across different specifications (see columns 1-8 of Table 3). The various specifications control for the different determinants of labour productivity growth. In the general model where we include all control variables (column 8 of Table 3), the coefficient of real house prices (β = −0.018) is negative and statistically significant, meaning that a standard deviation increase in real house prices index within country (an increase of 33 point) leads to a 0.6 percentage point (33*0.018) decline in labour productivity growth within country in our sample of OECD economies over the period of study (1972-2019). This is approximately 26% of a standard deviation of labour productivity growth rate. This finding provides empirical support for the theoretical works that argue house price appreciation have led to impaired labour productivity in most developed countries (e.g., Maclennan et al. 2015; Maclennan et al. 2018; Maclennan et al. 2021; Pawson et al., 2021).

Table 3. Fixed effects regressions: labor productivity and real house prices.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Dependent variable: Labor productivity
Real house prices	−0.023***	−0.021***	−0.022***	−0.023***	−0.022***	−0.023***	−0.018***	−0.018***
	(−4.99)	(−4.69)	(−4.68)	(−5.88)	(−4.91)	(−5.11)	(−3.37)	(−4.08)
KOF globalization		0.145***						0.113***
		(3.10)						(3.21)
Tertiary education			0.069***					0.081***
			(2.92)					(2.91)
Investment				0.063***				0.063***
				(2.97)				(3.18)
Democracy					0.014			0.014**
					(1.64)			(2.50)
Property rights						0.055***		0.058***
						(3.04)		(3.50)
Working age population							0.885*	0.665*
							(2.02)	(1.92)
Observations	826	826	826	826	826	826	826	826
R2 (within)	0.08	0.09	0.09	0.14	0.09	0.09	0.10	0.18

Notes: * 0.10, ** 0.05, *** 0.01. Robust t statistics are in parentheses and are based on clustered standard errors at country level. Country fixed effects are included in all models. Sample period: 1972–2019.

All of the control variables have the expected signs and are statistically significant. We find that higher level of KOF globalization, higher rate of tertiary education, stronger growth in investment, higher level of democracy and property rights, as well as higher ratio of working age population are positively associated with labour productivity growth within countries.

4.2. Robustness check: Panel ARDL-PMG estimation

Along with the panel fixed effect estimation, we also employ the panel ARDL-PMG estimator (Pesaran et al. 1997, 1999) which is suitable for our study because we are interested in understanding the long-run link between real house prices and labour productivity growth. The ARDL method is more appropriate for discovering long-term dynamics in the data, where both the time and cross-sectional elements are moderate to large, and the available frequency is annual (Baffes et al. 2022). In addition, this estimation method gives consistent parameter estimates in the existence of endogenous variables when the lag structure of the variables is correctly specified, irrespective of the order of integration of the variables (Pesaran et al. 1999). The ARDL method developed by Pesaran (1997) and Pesaran et al. (1999) has another attractive property, which examines long-run links regardless of the time-series characteristics of the individual regressors. Finally, the ARDL-PMG estimator is widely used to evaluate housing market dynamics (e.g., Duca et al. 2011; Gholipour et al. 2021).

Table 4 reports the estimation results. In columns 1 and 2, we examine the relationship between real house prices and labour productivity growth without and with control variables, respectively. The findings strongly confirm the adverse impact of higher real house prices on labour productivity growth as the coefficient of real house prices is statistically significant at the 1% level. For both specifications the short-run error correction coefficient is negative and highly significant, indicating that the variables show a return to long-run equilibrium.

Table 4. Results of panel ARDL/PMG estimations: labor productivity and real house prices.

Dependent variable: Labour productivity
Explanatory variables	(1)	(2)
	Without control variables	With control variables
Real house price	−0.024***	−0.020***
	(−8.549)	(−6.967)
Short-run error correction	−0.786***	−0.714***
	(−13.116)	(−10.591)
Period	1972-2019	1972-2019
Observations	999	997

Notes: *** 0.01. t statistics are in parentheses. Model selection method: Akaike info criterion (AIC). Maximum dependent lags: 1 (Automatic selection).

4.3. Sub-sample analysis

The full sample analyses provide robust evidence that there is a significant and negative relationship between real house prices and labour productivity growth. Next, we separate the full sample of countries into two categories based on their levels of economic complexity. The first group includes top nine economies that are highly complex and sophisticated: Japan, Germany, Switzerland, Sweden, Austria, Finland, United Kingdom, United States, and South Korea. The remaining economies are in the second group: France, Ireland, Italy, Belgium, Denmark, Israel, Netherlands, Spain, Canada, Norway, Portugal, New Zealand, Greece, Australia. Our categorization of countries are based on the average scores of Economic Complexity Index (ECI) over 1995-2021 developed by The Growth Lab at Harvard University (2019). Economic complexity is defined as “The economic complexity of a country is calculated based on the diversity of exports a country produces and their ubiquity, or the number of the countries able to produce them (and those countries’ complexity)⁵.”

We expect that the impact of house prices on labour productivity growth in highly complex economies is weaker than in less complex economies. This is because capital reallocation from other sectors to housing market during the housing boom occurs to a lesser extent in more complex economies than less complex economies. Therefore, lower capital investment would be diverted to the activities in the property market that are considered as low productive economic activities. This, in turn, reduces the adverse effect of housing boom on productivity.

The results of regressions for subsample countries are presented in columns 1–4 of Table 5. We find that house prices have a significant and negative impact on labor productivity growth in economies with lower level of economic sophistication such as Australia and Greece (columns 3 and 4 of Table 5). On the other hand, although the sign of real house price coefficient is negative for sub-sample of high economic complexity, but it is statistically insignificant (columns 1 and 2 of Table 5). These results imply that the association between house prices and labour productivity depends on the complexity of economies.

Table 5. Fixed effects regressions for subsamples - Higher economic complexity vs. Lower economic complexity.

	(1)	(2)	(3)	(4)
	Dependent variable: Labor productivity
	Higher economic complexity		Lower economic complexity
Real house price	−0.015	−0.010	−0.025***	−0.023***
	(−1.27)	(−1.85)	(−5.18)	(−4.79)
KOF globalization		0.111***		0.140**
		(3.88)		(2.54)
Tertiary education		0.163***		0.054**
		(5.19)		(2.31)
Investment		0.100***		0.050**
		(5.99)		(2.43)
Democracy		0.035***		0.008***
		(11.98)		(5.82)
Property rights		0.047**		0.066**
		(2.48)		(2.39)
Working age population		1.122**		0.121
		(3.05)		(0.33)
Observations	308	308	518	518
R2 (within)	0.03	0.30	0.11	0.16

Notes: **p < 0.05, ***p < 0.01. Robust t statistics are in parentheses. Country fixed effects are included in all models. The most complex economies, on average over 1995-2021, are included under “Higher Economic Complexity” sample. These 9 countries are: Japan, Germany, Switzerland, Sweden, Austria, Finland, United Kingdom, United States, and South Korea.

5. Conclusion

Labour productivity growth has sharply dropped in advanced economies since the late 1990s. The COVID-19 pandemic has seen labour productivity decelerate even further (Dieppe 2020). The slowdown has been attributed to various factors such as deceleration of the working-age population growth, stabilisation of educational attainment, loss in momentum of expansion into more varied and complex methods of production, deceleration reallocation within and between economic sectors, adverse shocks to productivity growth, slow adoption of information and communication technology (ICT) and declining contribution from ICT-intensive sectors in the US, and preventive product market rules in some parts of Europe (Dieppe 2020). Since the global financial crisis (GFC, 2007-2009), the key driver of slow labour productivity growth in developed economies is related to slowdown in capital deepening (Dieppe 2020).

In recent years, some scholars in both advanced countries and emerging economies have provided either firm-level evidence/or and theoretical arguments that the rises in housing market activities is another factor that should be blamed for slower labour productivity growth. In this study, we empirically explore the effect of increase in real house prices on labour productivity growth in a set of OECD countries over 1972-2019. By applying panel fixed effects and ARDL-PMG estimators and controlling for other determinants of labour productivity, we find that an increase in real house prices has a significant and negative impact on labour productivity growth in the long run across countries. We also find that the link between house prices and labour productivity is not homogenous across OECD countries and the significance and magnitude of the effect depends on the complexity of economies.

Overall, our findings suggest that government officials need to pay more attention to the impact of housing market dynamics on the labour productivity. One approach to diminish the adverse effect of house price increases on labour productivity is making the economy more diversified and complex. The complexity and sophistication in economies can possibly reduce the investment demand in housing markets, and slow down house price growth which has an adverse impact on labour productivity.

The findings of our research should be considered in light of its limitations. Due to data constraints, the present study only assesses the data of 24 OECD economies and thus generalisations of our results should be made with prudence. In addition, our estimations did not include the peak period of COVID-19 pandemic (2020-2022) when workers’ mobility was significantly limited and work from home became normal in most services industries. Future studies may repeat our study by considering the impact of pandemic on the link between house prices and labour productivity.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The data that support the findings of this study are available on request from the corresponding author.

Notes

Notes

1 See Rojas & Aramvareekul (2003), Chancellor (2015), Harrison (2007), Abdel-Wahab & Vogl (2011).

2 The causes of lower productivity growth in construction are construction’s poor total factor productivity (TFP) performance (Abdel-Wahab & Vogl 2011), inadequate research and development expenditures in construction industry, the classic labour-intensive nature of construction activities (Huang et al. 2009), and ongoing shortage of skilled workers (Huang et al. 2009).

3 Labour market competitiveness is Pillar 8 of the overall competitiveness index of World Economic Forum.

4 Based on the works of de Chaisemartin and D’Haultfoeuille (2020), Imai and Kim (2021), Kropko and Kubinec (2020), and Farzanegan et al. (2023), we have chosen not to include time fixed effects. This decision is based on the understanding that employing both country and time fixed effects can obscure the interpretation of the results.

5 See https://atlas.cid.harvard.edu/glossary