Spend less, get more? Explaining health spending and outcome differences between Canada and Italy

Di Matteo, Livio; Barbiero, T. P. Article Spend less, get more? Explaining health spending and outcome differences between Canada and Italy Review of Economic Analysis (REA) Provided in Cooperation with: International Centre for Economic Analysis (ICEA), Waterloo, Ontario Suggested Citation: Di Matteo, Livio; Barbiero, T. P. (2020) : Spend less, get more? Explaining health spending and outcome differences between Canada and Italy, Review of Economic Analysis (REA), ISSN 1973-3909, International Centre for Economic Analysis (ICEA), Waterloo (Ontario), Vol. 12, Iss. 4, pp. 403-438, https://doi.org/10.15353/rea.v12i4.1898 This Version is available at: https://hdl.handle.net/10419/328096 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc/4.0/ Review of Economic Analysis 12 (2020) 403-438 1973-3909/2020403 403 www.RofEA.org Spend Less, Get More? Explaining Health Spending and Outcome Differences Between Canada and Italy LIVIO DI MATTEO∗ Lakehead University THOMAS BARBIERO Ryerson University Canada spends more than Italy on health per capita and as a share of GDP and has a higher per capita GDP. Yet, life expectancy and infant mortality in Italy are better and have improved more over time. The implication is that the Italian health care system provides better value for money. We examine whether Italy does get better health outcomes at lower costs. Using regression analysis, we find that health spending is determined by similar drivers in both Canada and Italy. We also find that more social spending and health spending in either country do not satisfactorily explain the differences in health outcomes, suggesting the importance of broader socio-economic determinants like income and lifestyle choices. We conclude that while the levels of per capita health spending in Canada are higher than Italy, this partly reflects historical inertia in Canadian health spending partially attributable to the higher costs of health professionals relative to Italy. Keywords: health, expenditures,outcomes JEL classifications:I1, I3, F0, H1, N3 1 Introduction Canada and Italy are G7 countries with publicly funded universal health care systems that provide high quality health care. International rankings usually place Canada and Italy highly on health care performance. However, these rankings also often place Italy ahead in terms of health outcomes, even though when measured as a percentage of GDP and per capita expenditures Canada devotes more resources to health. For example, the WHO 20101 rankings placed Italy second and Canada 30th in terms of health system performance, while a more recent study in the Lancet 20182 placed Italy ninth and Canada fourteenth. Meanwhile, in terms of © 2020 Livio Di Matteo and Thomas Barbiero . Licensed under the Creative Commons Attribution - Noncommercial 4.0 License (http://creativecommons.org/licenses/by-nc/4.0/). Available at http://rofea.org. 1 World Health Organization (2010). 2 R. Lozano et al., (2018). Review of Economic Analysis 12 (2020) 403-438 www.RofEA.org 404 spending as a share of GDP, a recent report placed Canada third and Italy twenty fourth.3 Recent research shows that more health spending does not always result in improved outcomes4 and that “Expensive health care is not always the best health care.”5 The literature suggests that other social determinants also influence health outcomes. For example, Rowlingson (2011), Subramanian and Kawachi (2004) and Lynch et al., (2001), draw the link between income inequality and mortality rates as well as health and social problems. Dutton et al., (2018) for Canada emphasize social spending at the provincial government level as a determinant of improved performance when it comes to life expectancy. We compare health spending, health system parameters and outcomes in Canada and Italy to answer why Italy spends less on health care and yet has generally obtained better long-term outcomes as measured by life expectancy and infant mortality. This disparity in long-term outcomes is apart from the early differential performance of the two countries with respect to the impact of the Covid-19 pandemic which cannot be properly addressed until a longer-term perspective becomes available.6 The difference in long-term outcomes and performance may be rooted in the structure of the health and social expenditure network and support system in the two countries. Notably, Italy devotes much more to redistributive schemes compared to Canada, and this difference might explain its better long-term health outcomes. We also consider differences in basic health care inputs and institutions that may account for the disparity in health outcomes in the two countries. The efficacy and efficiency of the healthcare system is a pressing question in all advanced countries as they face the public health shocks from events such as the Covid-19 pandemic and aging populations that will put pressure on future government budgets. We also find that while Canada does spend more but get less given its level of spending, the health dollars spent at the margin seem to have yielded the same benefits in terms of life expectancy increases and infant mortality declines in both countries. The paper is divided as follows. First, we review the health care literature with respect to value for money in health care spending to provide context. Second, we overview the size, 3 Barua, B., S. Hasan and I. Timmermans (2017) Comparing Performance of Universal Health Care Countries, 2017. Fraser Institute. 4 Ariste and Di Matteo (2017). 5 OECD (2009) Health At a Glance. https://www.oecd.org/ health/expensivehealthcareisnotalwaysthebesthealthcaresaysoecdshealthataglance.htm 6 https://www.worldometers.info/coronavirus/#countries. Italy was hit early by the coronavirus and has less warning and thus, the Covid-19 pandemic in Italy was initially more pronounced than Canada, with dire consequences for the mortality rate. However, the situation continues to evolve and as of the end of July, total Italian deaths per one million population since the start of the pandemic exceed Canada’s total but on a current weekly basis, Italian deaths are now below Canadian. DI MATTEO, BARBIERO Health Spending and Outcomes in Canada and Italy www.RofEA.org 405 structure and financing of the two health care systems and compare aggregate health care spending in Canada and Italy, health care resource inputs, health care status indicators and outcomes, and social spending and income inequality. We then carry out a regression analysis to assess the determinants of health expenditures and use it to present a spending counterfactual simulation. We also estimate the determinants of health outcomes – namely, life expectancy at birth and infant mortality. The paper concludes with a discussion as to what factors help explain the differences in expenditures and health outcomes between Canada and Italy. 2 Value for Money in Health Care: A Brief Review of the Literature 2.1 Health Spending and Health Outcomes In a world of unlimited wants and scarce resources, efficiency, or getting value for money, is paramount. Getting the most out of inputs, such as diagnostic machinery, the number of doctors, and hospital beds, can translate in extending the length and quality of life. Input costs can be measured in monetary terms, such as the per capita spending on health care, or in physical terms, such as the number of hospital beds and technical equipment.7 Outcomes can be quantitative (lives saved or years of life gained), or qualitative (quality of additional years of life). A standard approach to efficiency in health economics is costeffectiveness analysis – that is a comparison of costs or expenditures per unit of health outcome.8 Health outcomes can be measured in terms of key indicators such as life expectancy at birth, mortality rates, or what are sometimes referred to as Quality Adjusted Life Years.9 Measures of health system cost-effectiveness and value for money could include health costs per quality adjusted life year, or measures of health spending per additional year of life expectancy or mortality rate reduction.10 7 There are also recent studies of health care system input costs at the international level and particularly the EU. See Joumard, André and Nicq (2010), Asandului, Roman and Fatulescu (2014), and Medeiros and Schwierz (2015). 8 See Folland, Goodman and Stano (2017: 81-107). 9 A quality adjusted life year is a valuation of a year of life that assigns values between 0 and 1 to represent quality of life in each year. As another example, the Canadian Institute for Health Information (CIHI) (2014) suggests another approach to measure a health care system: the reduction in potential years of life lost (PYLL) from treatable causes of death. According to their results, the reduction in PYLL in Canada from treatable causes of death could be anywhere between 18 and 35 percent, depending on the province, which translates to preventing between 12,600 and 24,500 premature deaths CIHI (2014: 2). In a separate study, the CIHI found that the PYLL for Italy for the year 2010 was lower than it was for Canada suggesting a more cost effective or efficient health care system given the greater per capita health spending in Canada CIHI (2016). 10 For a discussion of measurement of cost-effective and value for money, see Artiste and Di Matteo (2017). Review of Economic Analysis 12 (2020) 403-438 www.RofEA.org 406 The connection between expenditure on health and outcomes is not necessarily monotonic. Indeed, the research on the relationship between health spending and health outcomes often shows an ambiguous relationship. Some OECD countries, such as Spain, have high life expectancy, but with spending on health care in the middle of the pack. Belgium and Denmark rank high in spending but have only average health outcomes. Thus, we find in general a nonlinear relationship between health spending and outcomes, at least among European OECD member countries11. Babazono and Hillman (1994) analyzed 1988 data for OECD countries, including Italy, and show that total health care spending and outpatient and inpatient utilization rates of the health care system are not related to health outcomes. On the other hand, Anderson et. al (2000) look at health spending and outcomes for OECD countries for the 1960-1998 period and find a positive relationship between health care expenditures and health outcomes, particularly so with regards to the U.S.A. Golinelli et al., (2018) found mortality rates in Italy between 2011-14 increased because of government spending restraint, compared to the 1995 to 2010 period, implying that health expenditures do impact outcomes. However, citing Stuckler et. al (2010), they note that health spending cannot be considered the only determinant of mortality rates. In fact, Vercelli et. al. (2014) find mortality rates in Italy declining at a steady rate from the beginning of the 20th century, right up to 2008, while national income growth was more variable. Anderson and Frogner (2008) also find a tenuous relationship between health expenditures and health outcomes. They note that the U.S. had one of the highest per capita spending levels on health care in the world, yet it was equally likely to be at either the top or bottom of sixteen quality measures complied by the OECD. Other studies have also found a relatively small impact of health care spending. Among these are Filmer and Pritchett (1999), and Nolte and McKee (2004). As well, Nixon and Ulman (2006) emphasize the complexity of associating health expenditures with outcomes, and found that for 15 European nations, including Italy, between 1980 and 1995, increases in health care spending improved infant mortality, but only a marginal impact on life expectancy. Cremieux et. al. (1999) analyzed Canadian provincial data for the period 1978-1992 and found less spending on health care in Canada lead to an increase in infant mortality and life expectancy (LE). Namely, a 10% reduction in health care spending increased infant mortality by 0.5% and decreased LE by 6 months for men and 3 months for women. Cremieux et. al. (2005a, b) focused on the impact of spending on pharmaceuticals in Canada on health outcomes and found a strong positive relationship on the LE and infant mortality. On the other hand, 11 Madeiros and Schwierz (2015) DI MATTEO, BARBIERO Health Spending and Outcomes in Canada and Italy www.RofEA.org 407 Guindon and Contoyannis (2012) found no impact on health outcomes with increased spending on pharmaceuticals. Joumard et.al. (2010) take a more comprehensive input-output approach in comparing OECD countries relative efficiency. They find both the Canadian and Italian system to be quite efficient, but the Italian health care systems has a lower mortality rate. On an efficiency basis, Medeiros and Schwierz (2015) find that the Italian health care system is one of the top performers in terms of the amount spent on health care and outcomes, placing second on its constructed efficiency scale, just behind France, which spends more on health care. Thus, Italy is not far from the “efficiency frontier,” according to their analysis. 2.2 Social Spending, Income Inequality and Health Outcomes Health outcomes stem from a production function process in which health care or health spending is only one input, along with with additional socio-economic determinants.12 The broader determinants of health approach can include the role of schooling, spending on social programs and income inequality. Social spending is a large category that in the broadest sense is ultimately about income redistribution. It may include welfare payments due to various physical or mental challenges, employment insurance programs, affordable housing programs, tax breaks for social purposes or low-income families, and direct in-kind provisions of goods and services targeting disadvantaged groups. Some research shows social spending has important health effects on a population. On the topic of the role of social spending on infant mortality and birth outcomes, Kim and Saada (2013) find that income inequality and social policies, such as maternal leave policies may help to explain the cross-country variations in these important issues. Dutton et al., (2018) examine the association between health care spending, social programs and health outcomes in Canada using the ratio of social to health spending dollars. Using provincial data from 1981 to 2011 they find a 1-cent increase in social spending per dollar spent on health was associated with a 0.1% (95% confidence interval [CI] 0.04% to 0.16%) decrease in potentially avoidable mortality and a 0.01% (95% CI 0.01% to 0.02%) increase in life expectancy. However, this study used only provincial social spending excluding municipal and federal social spending and noted that the ratio of social spending to health spending across Canadian provinces was low. Bradley and Taylor (2013) found social spending in OECD countries improves health outcomes. They point out the stark example of the health care system in the U.S., one of the most expensive in the world, which ranks poorly on health outcomes. On a per capita basis, the 12 See Folland, Goodman and Stano (2017: 109-132). Review of Economic Analysis 12 (2020) 403-438 www.RofEA.org 408 U.S. spends double the OECD average on health, but the outcomes are among the worst in the OECD. Their research show that social service spending in the U.S. was predictive of three key health outcomes: life expectancy, infant mortality, and maternal mortality. Rubin et. al. (2016) make a similar point and show that old-age spending makes a significant contribution to better health outcomes. They suggest that social spending and protection may be more important to better health outcomes in societies in which income inequality is high. For the U.S., they find that unemployment and income maintenance are strongly correlated with better health outcomes across U.S. states, but social expenditures take time to translate into better health outcome. Bradley et. al. (2011) using a pooled cross section of OECD countries in 2009 find that health expenditures were significantly associated with better health outcomes in only two of five health indicators (life expectancy and maternal mortality) while social services expenditures adjusted for GDP were significantly associated with better health outcomes in three of five indicators. The ratio of social expenditures to health expenditures was significantly associated with better outcomes in infant mortality, life expectancy and increased potential life years lost, after adjusting for the level of health expenditures and GDP. Lynch et. al (2004) in a review of 98 aggregate multilevel studies found income inequality is not associated with health outcome differences in affluent countries. Despite the tenuous relation between income inequality and health outcomes, they do claim that reducing income inequality of the most disadvantaged populations will improve their health and thus the health of general population. Latif (2015) examined the effect of income inequality on health outcomes in Canada and found no significant impact, but that absolute household income in Canada had a significant positive impact on health status. Wilkinson and Pickett (2015) in a literature review of epidemiological causal literature find support for the proposition that that income inequality affects population health and wellbeing. Wilkinson and Pickett (2011) made similar claims in their previous work. Snowden (2011) in a direct response to the Wilson and Pickett position on income and inequality and population health dissents and argues that health outcomes are the product of diverse influences that over time can impact life expectancy and other health indicators. This paper contributes to the current literature by having a closer look at two of the G7 countries that devote different amounts to health spending, and with different social expenditures and lifestyle factors to determine the contributions to health outcomes. We want to gain insights on whether Canada or Italy is getting greater relative value for money. DI MATTEO, BARBIERO Health Spending and Outcomes in Canada and Italy www.RofEA.org 409 3 Health System Overview: A Comparison of Health Expenditure, Indicators and Outcomes, and System Trends 3.1 Canada’s Health System In Canada, federal and provincial governments finance public health spending but under the Canadian constitution the provincial governments deliver publicly funded health care to citizens, making for a decentralized health system with some variation in both expenditure levels and service provision as well as some reliance on market mechanisms for services.13 Indeed, the OECD puts Canada into Group 2 countries along with Belgium, Australia and France - countries with public insurance for basic coverage and private insurance beyond the basic coverage.14 Canadian public health care is not one public health care system, but is 14 publicly funded systems, given that there are ten provinces and three territories along with a federal government with its own health obligations for indigenous peoples, the military and the RCMP. Provincial and territorial government health expenditures are for insured health services and extended health care and are financed by own source revenues as well as federal government transfers to the provinces. Canada’s public health care system evolved in stages, starting with individual provincial initiatives that were followed by federal financing. In 1957, the federal Hospital Insurance and Diagnostics Act established cost-sharing for provincial hospital insurance plans while the 1966 federal Medical Care Insurance Act provides cost sharing for physician services also. Thus, health system financing in Canada involves several layers: 1) public services for medically necessary hospital, diagnostic and physician services, 2) a mix of public-private insurance covered services including drugs and home care and 3) entirely private services such as dental services and physiotherapy.15 Canadian health care is approximately 30 percent privately and 70 percent publicly funded, but the proportion varies across provinces, as well as within expenditure categories.16 Private sector health care expenditures include those from health insurance firms, out-of-pocket expenditures of individuals, and patient service revenue paid by private insurers for items such as preferences for private hospital rooms or charges for services not deemed medically necessary. It also should be noted that in Canada public finance differs from public provision. 13 Marchildon (2019). 14 OECD 2010, “Health care systems: Getting more value for money”, OECD Economics Department Policy Notes, No.2. 15 Martin et al. (2018) 16 See: Canadian Institute for Health Information. National Health Expenditure Trends 2018, 1975 to 2017. Review of Economic Analysis 12 (2020) 403-438 www.RofEA.org 410 Canadian physician expenditures are nearly entirely publicly financed with physicians behaving mainly as private independent contractors rather than salaried employees.17 Physicians in Canada have substantial autonomy and flexibility in their daily practices and access to physician services can sometimes be problematic with reliance on hospital emergency wards or walk-in clinics.18 Federal transfers for provincial-territorial government health spending are made under the 1984 Canada Health Act, which specifies the criteria under which provinces and territories get federal health transfers. Provincial health systems must be publicly administered, comprehensive in their coverage of insured services, universal in their coverage, must have portable benefits, and must provide reasonable access to insured services. There is some variation across provincial systems in terms of what medical services and drugs are covered by provincial plans, but health outcomes are remarkably uniform across provinces.19 In 2018-2019, the total value of the Canada Health Transfer to the provincial and territorial governments according to the Federal Department of Finance is expected to be $CAD 38.584 billion dollars and is expected to increase to $CAD 40.373 billion by 2019-2020.20 These federal transfers provide about 20 to 25 percent of provincial government health expenditures, though the growth rate of these transfers changed after 2017 as a new transfer formula took effect. The growth rate of the Canada Health Transfer is now based on the growth rate of real GDP, subject to a 3 percent floor as opposed to the previous Health Accord escalator, which saw annual 6 percent increases.21 3.2 Italy’s Health system The OECD groups the Italian health care system in Group 6, along with those of Hungary, Ireland, New Zealand, Norway, Poland and the U.K - health care services provided by a heavily regulated public centralized system, with budgetary spending limits. 22 Italy’s publicly funded health care system (Servizio Sanitario Nazionale-SSN) was established in 1978 and modelled on Britain’s National Health Service. It is a single payer system providing universal coverage for all Italians, with responsibility for the organization and delivery of health services residing with the 19 regions and two autonomous provinces. 17 For a discussion see Di Matteo (2014). 18 See Di Matteo (2014) and Clarke (2016). 19 See Marchildon (2019) for further discussion. 20 See: Department of Finance Canada. https://www.fin.gc.ca/fedprov/mtp-eng.asp 21 Norris (2018). 22 OECD 2010, “Health care systems: Getting more value for money”, OECD Economics Department Policy Notes, No. 2. Canada in in Group 2, which also includes Australia, Belgium and France. DI MATTEO, BARBIERO Health Spending and Outcomes in Canada and Italy www.RofEA.org 417 in 2007 and has since remained above. Between the 1960s and the 2010s, average annual life expectancy at birth in Canada rose from 71.7 to 81.7 years, while in Italy it rose from 69.8 to 82.7 years. Of additional interest, suicide rates (intentional self-harm) in both Canada and Italy are below the OECD average, and both have declined since the mid-1980s with suicide rates in Italy dramatically lower than in Canada. We use aggregated social expenditure data from the OECD. 29 As a share of GDP, social expenditure has risen over time in both Italy and Canada, but Italy generally spends more (See Figure 2). For example, in 1980 Canada spent 13.3 percent of GDP on social programs, compared to 17.4 percent for Italy. Over the last approximately 40 years, that percentage has grown to 28.1 percent in Italy in 2017 and ultimately reached 17.3 percent in Canada after declining from a peak of 18.4 percent in 1995 in the wake of the intergovernmental transfer cuts of the federal fiscal crisis. Figure 2: Social Expenditure as a Percent Share of GDP Source: OECD Stat. https://stats.oecd.org/Index.aspx?datasetcode=SOCX_AGG 29 OECD Stat. https://stats.oecd.org/Index.aspx?datasetcode=SOCX_AGG Review of Economic Analysis 12 (2020) 403-438 www.RofEA.org 418 There are differences in social expenditure focus in the two systems. Public expenditures on old age and survivors cash benefits are higher in Italy, reflecting a more generous Italian social security system. In 1980 Canada spent 3.1 percent of its GDP on old-age related transfers, and this grew to 4.7 percent in 2015. In Italy it already reached 8.6 percent in 1980, and by 2015 it almost doubled to 16.2 percent. By comparison, the OECD average spending on old-age related transfers was 5.5 percent of GDP in 1980 and 7.5 percent in 2015. Public expenditures on disability and sickness cash benefits are nearly double in Italy compared to Canada, at 0.8 percent of GDP in Canada and 1.7 percent in Italy in 2015. Direct public expenditures on family (in cash and in kind) were 1.5 percent of GDP in Canada, and 2 percent of GDP in Italy in 2015. Telling of the extent of the redistribution of income through social programs is the income of older people aged 65 and over, as a percent of the general population income. In 2014 those 65 years and older in Canada had 91.1 percent of the income of the general population, as compared with 98.8 for Italy, while the OECD average was 87.6 percent. Figure 3: Top 1% Percent Share of Adult Pre-Tax Income Source: World Inequality database https://wid.world/ DI MATTEO, BARBIERO Health Spending and Outcomes in Canada and Italy www.RofEA.org 419 With respect to income inequality, OECD Gini coefficients30 for Canada and Italy show that in 2016 Italy was slightly more unequal than Canada with a Gini of 0.33, compared to 0.31 for Canada. However, measuring inequality as the pre-tax income share of the top 1% of adult earners, Canada appears to have greater income inequality than Italy. The income inequality data for Italy should be treated cautiously, since it has a relatively large shadow economy,31 making officially reported market income to construct inequality measures potentially unreliable. Figure 3 plots the pre-tax income share of the top 1 % of adult earners for Canada and Italy using data from the World Inequality Database.32 The top 1 percent of earners was used as the inequality measure given that more preferable Gini coefficient data was not available. While income inequality as measured by the top 1 percent of earners has grown in both countries over time, Canada has greater income inequality than Italy, at least before taxes and transfers are factored in. As noted, the relatively large size of the black-market economy in Italy compared to Canada means that these data may understate the extent of regional or individual level inequality. To summarize, Canada spends more than Italy on health care, but Italy has surpassed Canada in basic health indicators like longevity and infant mortality. Despite spending more on health both per capita and relative to GDP, Canada per capita has fewer physicians, fewer hospital beds and less medical technology, but does have more nurses and physician consultations. Despite spending more, based on the basic health indicators of longevity and infant mortality Canada appears to be getting less. 4 Methodology We conduct regression analysis on health care spending drivers and key health indicators to determine the responsiveness of health spending and health indicators to common determinants for Canada and Italy. Along with using the results to examine the determinants of health spending, the regression results are also used to conduct a counterfactual simulation. 4.1 Regression Analysis There is an extensive literature on the determinants of health spending – both public and private – with key expenditure drivers including population growth, physician numbers given the gate keeping role of physicians,33 population aging, income, inflation, and enrichment factors such 30 https://data.oecd.org/inequality/income-inequality.htm 31 See Medina and Schneider, 2018 for a discussion. 32 https://wid.world/ 33 For a discussion see Di Matteo (2014). Review of Economic Analysis 12 (2020) 403-438 www.RofEA.org 420 as technological change as often by diagnostic equipment usage or by time trend. These studies have been carried out at the national, international and regional levels.34 Even though Canada and Italy are very regionalized countries, we focus on national level determinants of health expenditures for several compelling reasons. First, while Canadian health care is delivered by the provinces with regional differences in per capita spending and services, health outcomes are consistently high across the country. Italy also has consistency in health outcomes across its regions, though there are regional differences in spending (Francese and Romanelli 2011)35. Second, the lack of available regional level data across a long enough time span for both spending, service provision and health outcomes make it impossible for our study to focus on regions. While sufficient and comparable time series going back to 1970 are either available or can be constructed at the national level, the data issue becomes more problematic at the region/province level. For Canada, provincial level health expenditure data is available going back to the mid-1970s from the National Health Expenditure Database of the Canadian Institute for Health Information. For Italy, the data sets for regional spending only go back to the 1990s. Third, a high-level national comparison is better given institutional differences in the roles of regions/provinces in health care across Canada and Italy. Canadian provinces have considerably more autonomy in the areas of health and education spending than do Italian regions. Canada is a decentralized federation where the provinces deliver and fund health care, whereas in Italy, regional health care provision is accompanied by much more centralization with respect to funding and provision. Thus, our decision to rely on national level data. We model the determinants of health expenditure in Canada and Italy as: Ht = f(z1t, z2t, ….znt) (1) where Ht is real per capita government health expenditures and z1 to zn represent a vector of social, demographic, economic and policy variables at time t which are determinants of Ht.36 34 See Constant et al., (2011). For an excellent survey of the international health expenditure determinants literature, see Gerdtham and Jonsson (2000). The first generation of such determinants studies often used international data. See Leu (1986), Parkin et. al., (1987), and Gerdtham et al (1992). See also Hitiris and Posnett (1992), Barros (1998), Gerdtham et. al, (1998), Di Matteo and Di Matteo (1998), Ariste and Carr (2003) and Crivelli, Filippini and Mosca (2006). For some more recent papers, see Cantarero Prieto and Lago Penas (2010), Magazinno and Mele (2012), Di Matteo (2010, 2014), Di Matteo and Emery (2014), Braendle and Columbier (2016), Nghiem and Connelly (2017), Yetim et al, (2020). 35 Generally, the southern part of Italy has lower per capita healthcare spending. The regional differences in healthcare spending is partly offset by southern Italians travelling north for health care services. 36 The variables are: PPP GDP Per Capita Current Prices; Population (millions); Unemployment Rate (%); Total Health Exp to GDP (%) HGDP; Per Capita Total Health Expenditure USPPP$; Public DI MATTEO, BARBIERO Health Spending and Outcomes in Canada and Italy www.RofEA.org 421 The national level data for these regressions covers the period 1970 to 2017, and comes from OECD Health Statistics 201837, supplemented by data from the World Bank, the OECD Social Health Expenditure Database38, the IMF WEO Database39, the World Inequality Database,40 and FRED.41 The specification is log-linear, and the estimation technique is Robust Regression which helps address the potential impact of outliers in data.42 We conducted Augmented Dickey-Fuller tests on the regression variables and found a high degree of stationarity across the variables in the dataset.43 Table 4 reports our results for per capita public and per capita total health expenditures. The final reported specifications regress the health expenditure variables on GDP per capita in U.S. PPP dollars lagged one year44 (a measure of long term income impact), the first difference of GDP per capita in U.S. PPP dollars (a measure of short term income fluctuations), physicians per 1,000 population, nurses per 1,000 population, the number of MRI machines per one million population (a variable to represent the impact of diagnostic technology on health spending), and the percent of population aged 65 years and over using a quadratic specification to reflect the fact that while aging populations can drive up health spending, eventually steps will be taken to restrain it making for a hump shape. We find differences driving per capita health spending across the two countries. Lagged per capita GDP is a significant long-term driver of per capita Canadian health spending, but shortterm fluctuations in GDP are not, with opposite response to these two variables for Italy. It is unclear why per capita health spending is less responsive to the level of GDP in Italy compared to Canada, but given the significance of the first difference of GDP per capita for Italy (but not Health Exp Per Capita USPPP$; Infant Mortality Deaths per 1000 live births; Life Expectancy at Birth Years; Self Harm (Suicides) Deaths per 1000; Alcohol Consumption Litres Per Capita; Tobacco Consumption. % pop 15+ who smoke; Physicians per 1000 Population; Medical Graduates Per 1000 Population; Nurses per 1000 population; Total Hospital Beds per 1000 population; MRIs per one million people; CTs per one million people; Doctor consultations Per Capita; Percent of population Aged 65 Years and Over; Social Expenditure Share of GDP (%); Social Expenditure Per Capita USPPP$; Top 1 Percent Pre-Tax Income Share. 37 http://www.oecd.org/els/health-systems/health-data.htm 38 http://www.oecd.org/social/expenditure.htm 39 https://www.imf.org/external/pubs/ft/weo/2018/02/weodata/index.aspx 40 https://wid.world/ 41 FRED is the Federal Reserve Bank of St. Louis Database. See: https://fred.stlouisfed.org/ 42 It should be noted that while the data for these regressions is from 1970 to 2017, some of the variables did not have complete coverage and as a result gaps were filled using a regression that regressed the available data on time trend and the coefficients used to fit the data to the missing years. 43 The ADF test was done employing no constant or trend and two lags. For most of the variables, the hypothesis of a unit root was rejected at the 5% and 10% level and occasionally at the 1% level. 44 Initial specifications also included 2- and 3-year lags, but they were found to be of low impact. Review of Economic Analysis 12 (2020) 403-438 www.RofEA.org 422 Table 4: Health Expenditure Determinant Regression Results Dependent Variable: Log of Per Capita Total Health Expenditure USPPP$ Canada Italy Coefficient t-statistic Coefficient t-statistic GDP Per Capita in PPP Dollars Lagged One Year 0.000039 4.06 0.000004 0.33 First Difference of GDP Per Capita in PPP Dollars 0.000004 0.24 0.000039 1.91 Physicians per 1000 Populatoin 0.498666 3.83 0.648925 2.32 Nurses per 1000 Population 0.002573 0.10 1.187990 9.88 MRI per Million Population -0.033041 -1.9 0.010492 1.25 Percent of Population Aged 65 Years and Over 0.749569 7.23 -0.014850 -3.81 Percent of Population Aged 65 Years and Over Squared -0.026910 -8.42 -0.014850 -1.89 Constant 0.607742 1.00 0.903599 0.81 R-squared 0.9028 0.7639 F-Statistic (7,39) 1349.63 854.35 Dependent Variable: Log of Public Health Expenditure Per Capita USPPP$ GDP Per Capita in PPP Dollars La gg ed One Yea r 0.000043 3.72 0.000014 0.77 First Difference of GDP Per Capita in PPP Dollars 0.000001 0.08 0.000077 2.43 Ph y sicians per 1000 Populatoin 0.692374 4.44 0.545579 1.27 Nurses per 1000 Populatio n 0.000212 0.01 1.110165 6.00 MRI per Million Population -0.042219 -2.03 0.021116 1.64 Percent of Population A g ed 65 Years and Ove r 0.626672 5.05 0.249538 1.31 Percent of Population A g ed 65 Years and Over Square d -0.023560 -6.16 -0.016379 -2.73 Constan t 0.8083971.12 1.137162 0.67 Ad j usted R-square d 0.8962 0.8270 F-Statistic (7,39) 869.39 321.66 Bold denotes significant at 5 percent level; Bold italics is significance at 10 percent level for Canada) it may be that there is more sensitivity in Italian health expenditures to changing economic circumstances. Italian health spending is more centralized than Canada which means that changes resulting from resource fluctuations can be transmitted more rapidly across the system.45 45 GDP per capita has gone up a bit more in Canada compared to Italy since 1990. In 1990 GDP per capita in Italy was US$18.55K and in Canada US$20.25k. In 2016 it was 44.92k in Canada and 39.04 in Italy: https://data.oecd.org/gdp/gross-domestic-product-gdp.htm. It should be noted that the data is PPP at current prices. DI MATTEO, BARBIERO Health Spending and Outcomes in Canada and Italy www.RofEA.org 423 With respect to the impact of GDP on health spending, it is important to note Italy’s shadow economy is generally held to be larger than Canada’s. According to a study by the University of Tubingen in West Germany (IAW), as a percentage of GDP Italy’s shadow economy was almost 20 percent (19.8%), while Canada’s shadow economy in 2017 is estimated at 9.8 percent.46 The relatively large shadow economy in Italy probably understate levels of GDP and GDP per capita. It should also be emphasized, that if GDP and GDP per capita are understated it could make the Italian health care system look more efficient, since it implies that Italy spends a lower percentage of its GDP on health care. Physician numbers are positive drivers of spending in both Canada and Italy and the coefficient estimates suggest the response is weaker for Canada when it comes to total health spending, but greater for public health spending. We note that the larger coefficient on physician numbers in Canada as a determinant of public health spending relative to Italy may be due to family physicians being paid on a capitation basis in Italy, whereas in Canada the feefor-service approach still dominates. It is also possibly a function of physicians paid relatively more in Canada than Italy. In 2015 Canadian specialist physicians average gross income was 4.5 times more than the average wage while in Italy the ratio was 2.5.47 As another example, in 2011, specialist physicians in Canada in US$ PPP received annual remuneration of $213,000 while in Italy it was $93,000.48 The lower remuneration for Italian physicians partly enables a relatively larger number of medical doctors per 1000 inhabitants: In 2018 Italy had 3.9 doctors while Canada had 2.5.49 The number of doctors grew very rapidly in the 1970s and 1980s (Walton and Binns, 1984) in Italy, which put downward pressure on doctors’ salaries. The rapid rise in the number of doctors in Italy in the 1970s and 1980s is attributable to the elimination of entry barriers in medical schools in the late 1960s.50 Students wishing to study medicine could enroll in the faculty of medicine of any Italian university. However, the same explanation does not appear to hold for nurses. Nurses per 1,000 population is not a significant driver of Canadian health spending at either the total or public 46 Reported in Forbes, February 9, 2017, https://www.forbes.com/sites/niallmccarthy/2017/02/09/wherethe-worlds-shadow-economies-are-firmly-established-infographic/#7459be0c742c 47 Source: OECD (2017), “Remuneration of doctors (general practitioners and specialists)”, in Health at a Glance 2017: OECD Indicators, OECD Publishing, Paris. 48 Remuneration of Doctors and Nurses: Progress and Persisting Issues. Joint Sessions of Health Data correspondents and Health Accounts Experts. Paris, 17 October 2013. OECD. https://www.oecd.org/els/health-systems/Item6_Remuneration-of-doctors-and-nurses_Durand.pdf. 49 Italy has one of the highest numbers of medical doctors in the OECD. There are historical reasons for this. In the late 1960s access to medical schools was made available to anyone that wanted to register for them and led to an explosion of graduates. See Walton and Binns (1984). For many years in the 1980s and 1990s there were many unemployed medical doctors in Italy, which ultimately had an impact on doctors’ salaries. 50 Thorne (1996). Review of Economic Analysis 12 (2020) 403-438 www.RofEA.org 424 level but is for Italy. This is an intriguing result given Italy has more physicians but fewer nurses per capita than Canada, and Canadian nurses are paid substantially more than Italian nurses. Again, according to the OECD the remuneration of hospital nurses in US$PPP in 2011 was $51,000 in Canada and $37,000 in Italy.51 This suggests that it is not numbers or costs alone, but perhaps differences in use and role of nurses in the health care process. New diagnostic technology has a differential impact on spending across the two countries. For Canada, increasing the number of MRI units is associated with a decrease in per capita health spending at both the total and public level. For Italy, the result has been positive, though only significant for public health care spending and at the 10% level. Again, this suggests a potentially differential effect across Italy and Canada over time on per capita health spending from the introduction of new diagnostic technologies. Technological change can have costenriching effects if increased demand results from new procedures. Even if the new techniques generate less expensive health procedures, the ultimate impact on spending will depend on the interplay of the two forces.52 Finally, an aging population is a positive driver of health spending in both countries, but based on the coefficient sizes, the magnitude of the effect is greater in Canada. In both countries, the relationship between proportion of population aged 65 and over is a quadratic (hump-shaped). With respect to total per capita health spending, an aging population has its peak positive effect on health spending in Canada until the percent aged 65-and-over reaches 14 percent of the population, and for Italy when that percentage reaches 8 percent. This suggests that Italy may have moved earlier to deal with the costs of an aging population. However, it could also be a “statistical illusion” in that the over-65 population in Italy began to grow at an earlier period compared to Canada. The efficacy of both health and social spending on life expectancy (LE) and infant mortality is considered in the regressions in Table 5. The log of life expectancy at birth and the log of infant mortality rates are regressed on the following key variables: lagged level of per capita GDP in U.S. PPP dollars, alcohol consumption per capita (litres), tobacco consumption53, 51 Source: Remuneration of Doctors and Nurses: Progress and Persisting Issues. Joint Sessions of Health Data correspondents and Health Accounts Experts. Paris, 17 October 2013. OECD. https://www.oecd.org/els/health-systems/Item6_Remuneration-of-doctors-and-nurses_Durand.pdf. 52 For example, Cutler et al. (1998) report that the real quality-adjusted price of heartattack treatments declined at an annual rate of 1.1 percent between 1983 and 1994. 53 It should be noted that alcohol consumption between males and females on an annual basis for the period under review was not available. As for tobacco consumption, while separate data series for males and females were available, both have trended down over time and since second-hand smoke can be just as dangerous as first-hand smoke, we opted to use the aggregate data. DI MATTEO, BARBIERO Health Spending and Outcomes in Canada and Italy www.RofEA.org 425 Table 5: Health Status Indicators Determinants Regression Results Dependent Variable: Log of Life Expectancy at Birth in Years Canada Italy Coefficient t-statistic Coefficient t-statistic GDP Per Capita in PPP Dollars Lagged One Year 0.000002 9.81 0.000003 6.80 Alcohol Consumption (Litres per capita) -0.002733 -5.35 -0.000003 0.00 Tobacco Consumption (%15 years plus who smoke) -0.000538 -1.64 -0.003852 -4.62 First Difference of Per Capita total Health Expenditure in PPP Dollars 0.000006 0.56 0.000000 0.03 First Difference of Social Expenditure Per Capita in PPP Dollars -0.000002 -0.48 -0.000004 -0.49 Total 1% Income Share -0.001271 -2.43 0.000151 0.10 Constant 4.353289 325.94 4.404139 139.96 R-squared 0.853000 0.875700 F-Statistic (6,40) 1238.18 764.73 Dependent Variable: Log of Infant Mortality (Deaths per 1000 Live Births) GDP Per Capita in PPP Dollars Lagged One Year -0.000016 -1.83 -0.000048 -8.15 Alcohol Consumption (Litres per capita) 0.137729 7.01 0.027815 2.59 Tobacco Consumption (%15 years plus who smoke) 0.013981 1.11 0.012768 0.97 First Difference of Per Capita total Health Expenditure in PPP Dollars -0.000633 -1.60 -0.000047 -0.20 First Difference of Social Expenditure Per Capita in PPP Dollars 0.000015 0.08 -0.000102 -0.78 Total 1% Income Share 0.051073 2.54 -0.018311 -0.81 Constant 0.241695 0.47 2.454694 4.94 R-squared 0.8109 0.8654 F-Statistic (6,40) 100.02 769.09 Bold denotes significant at 5 percent level; Bold italics is significance at 10 percent level defined as the percent of population aged 15 years and over who smoke, the first difference of per capita total health expenditure in U.S. PPP$, the first difference of social expenditures per capita in U.S. PPP$ 54 and the income share of the top 1 percent. We use per capita GDP rather than changes in GDP as the determinant of changes in health status given its standard use in the 54 Initial data exploration using LOWESS found that the relationship between life expectancy and the per capita health spending variable or its first difference was increasing. Similarly, the LOWESS plot between infant and per capita health spending or its first difference was a decreasing relationship. As a result, quadratic specifications were not employed. Review of Economic Analysis 12 (2020) 403-438 www.RofEA.org 426 literature.55 With respect to public spending, changes at the margin are more immediate input factors affecting current health status – therefore, the use of first differences in health and social expenditure.56 As in our first set of regressions, Robust regression estimates were done. Table 5 suggests that for both countries, the level of per capita GDP is a key determinant of increased life expectancy and falling infant mortality for Italy, but not for Canada. Tobacco consumption significantly reduces life expectancy in both countries (but at the 10 percent level for Canada), while alcohol is correlated with reduced life expectancy in Canada but not Italy. Alcohol consumption significantly increases infant mortality in Italy and Canada, while tobacco consumption increases infant mortality in Canada and Italy but is not statistically significant in either country. The health effects of tobacco consumption during pregnancy have been widely known for decades, and the reductions in consumption rates in both countries over time may be a factor in the insignificance of the results with respect to infant mortality. The effect of marginal changes in total health spending on health outcomes are insignificant drivers of life expectancy and infant mortality in both countries. Moreover, when it comes to social spending, changes in expenditures at the margin are also insignificant contributors to life expectancy and infant mortality. Given the high level of social and economic development of Canada and Italy, it is not surprising that changes in health and social spending at the margin are not significant determinants of health status. Instead, the key drivers appear to be the longterm level of economic development as captured by levels and changes in income, with some influence of tobacco and alcohol use. As for income distribution and inequality, we find that an increase in the income share of the top 1% in Canada is statistically significant in decreasing life expectancy and increasing infant mortality. For Italy, an increase in income inequality is not a statistically significant determinant of either life expectancy or infant mortality. While it may be that the effects of income inequality on health outcomes may be weaker in some developed countries with good public health systems, the Canadian results may be driven by the poorer health outcomes among its indigenous peoples.57 Another possibility driving Italian results is the large proportion of social transfers devoted to income redistribution, especially for the elderly, which may serve to weaken the effect of market income inequality on life expectancy. We have already noted that in Italy market incomes do not reflect the broader underground economy, which implies that official income figures used also to estimate inequality do not reflect the true extent of either individual or regional inequality in Italy. 55 The health determinants literature focuses on level of income as a key determinant of health spending. 56 The levels of these two expenditure variables are also highly correlated with GDP. 57 For discussion, see Ring and Brown (2003) and Smylie and Firestone (2016). It should be noted that it would have been superior to use the Gini coefficients in both countries, but consistent annual data is not available. DI MATTEO, BARBIERO Health Spending and Outcomes in Canada and Italy www.RofEA.org 433 Canadian Institute for Health Information (2018). National Health Expenditure Trends 2018, 1975 to 2017. Canadian Institute for Health Information (2016). Canada’s International Health System Performance Over 50 Years: Examining Potential Years of Life Lost. Ottawa, ON: CIHI Canadian Institute for Health Information (2014), “Health System Efficiency in Canada: Why Does Efficiency Vary Among Regions?” Ottawa, ON: CIHI Canadian Institute for Health Information. (2012). Developing a Model for Measuring the Efficiency of the Health System in Canada. Ottawa, ON: CIHI Cantarero Prieto, D. and Lago Peñas, S. (2010): “The determinants of health careexpenditure in Spain: a re-examination”, Applied Economics Letters, 17(7-9), 723-726. Cavalieri, M. Guccio, C. (2006), Health Expenditure in Italy: A Regional Analysis of the Public-Private Mix, Dipartimento di Economia Pubblica e Territoriale, University di Pavia https://pdfs.semanticscholar.org/f514/f83059b503625cb28776f6392da19ed38a71.pdf Cicchetti, A, Gasbarrini, A (2016), “The healthcare service in Italy: regional variability,” European Review for Medical and Pharmacological Science, Vol 20 (1 suppl), 1-3 Clarke, J. (2016) Difficulty accessing health care services in Canada. Health at a Glance. Statistics Canada. Catalogue no.82-624-X. https://www150.statcan.gc.ca/n1/en/pub/82- 624-x/2016001/article/14683-eng.pdf?st=oWNTo5Jc Constant, A., S. Peterson, C.D. Mallory and J. Major. 2011. Research Synthesis on cost Drivers in the Health Sector and Proposed Policy Options. Ottawa, ON: Canadian Health Services Research Foundation, Reports on Cost Drivers and Health System Efficiency: Paper 1. Cremieux, P. Y., Meilleur, M. C., Ouellette, P., Petit, P., Zelder, M., & Potvin, K. (2005a). Public and private pharmaceutical spending as determinants of health outcomes in Canada, Health Economics, 14(2), 107–116. Cremieux, P. Y., Ouellette, P., & Pilon, C. (1999). Health care spending as determinants of health outcomes. Health Economics, 8, 627–639. Crivellei, L., M. Filippini and I. Mosca (2006): “Federalism and regional healthcare expenditures: an empirical analysis for the Swiss cantons,” Health Economics,15(5), 535- 541. Cutler, D.M., et al. 1998. “Are Medical Prices Declining? Evidence from Heart Attack Treatments.” Quarterly Journal of Economics, 113(4): 991-1024. Di Matteo, L. (2014) “Physician numbers as a driver of provincial government health spending in Canadian health policy,” Health Policy, 115, 1, 18-35. Di Matteo, L. (2010). “The sustainability of public health expenditures: evidence from the Canadian federation.” European Journal of Health Economics, 11 (6): 569-584. Review of Economic Analysis 12 (2020) 403-438 www.RofEA.org 434 Di Matteo, L. and R. Di Matteo. (1998). “Evidence on the Determinants of Canadian Provincial Government Health Expenditures: 1965-1991.” Journal of Health Economics 17(2): 211- 228. Di Matteo, L. and J.C.H. Emery (2014) “Common provincial determinants and cost drivers,” Chapter 4, co-authored with J.C.H. Emery, in Bending the Cost Curve in Health Care: Canada’s Provinces in International Perspective, eds. G.P. Marchildon and L. Di Matteo. The Johnson-Shoyama Series on Public Policy, University of Toronto Press. Dutton, D.J., P-G Forest, R.D. Kneebone, J.D. Zwicker (2018) “Effect of provincial spending on social services and health care on health outcomes in Canada: an observational longitudinal study” CMAJ, January 22nd, Vol. 190, No. 3, E66-E71. Ferré F, de Belvis AG, Valerio L, Longhi S, Lazzari A, Fattore G, Ricciardi W MA.(2014). Italy: Health System Review. Health Systems in Transition, 2014. 2014. Available from: http://www.euro.who.int/__data/assets/pdf_file/0003/263253/HiT-Italy.pdf?ua=1 Filmer, D.,&Pritchett, L. (1999). The impact of public spending on health: Does money matter? Social Science and Medicine, 49, 1309–1323. Folland, S., A.C. Goodman and M. Stano (2017) The Economics of Health and Health Care, 8th edition, Routledge: New York. Francese, M. and Romanelli, M. (2011). “Healthcare in Italy: Expenditure Determinants and Regional Differentials,” Banca D’Italia Working Papers, No. 828 Gerdtham, U.G., J. Sogaard, F. Andersson and B. Jonsson. (1992). “An econometric analysis of health care expenditure: A cross-section study of the OECD countries,” Journal of Health Economics 11: 63-84. Gerdtham, U.G., B. Jonsson, M. MacFarlan and H. Oxley. (1998). “The Determinants of Health Expenditure in the OECD Countries: A Pooled Data Analysis,” P. Zweifel, ed. Health, The Medical Profession and Regulation. Developments in Health Economics and Public Policy 6: 113-134. Gerdtham, U.G. and B. Jonsson. (2000). “International Comparisons of Health Expenditure: Theory, Data and Econometric Analysis,” Handbook of Health Economics, Vol. 1, eds. A.J. Culyer & J.P. Newhouse. Elsevier Science. Golinelli, D., Bucci, A., Toscano, F., Filicolori, F., and Fantini, P (2018), BMC Health Services Research 18:671 Gori, C. (2012). Home care in Italy: a system on the move, in the opposite direction of what we expect. Special Issue: Reforming Home Care in Ageing Societies, Vol. 20, Issue 3. https://doi.org/10.1111/j.1365-2524.2011.01052.x Guindon, G. E., & Contoyannis, P. (2012). A second look at pharmaceutical spending as determinants of health outcomes in Canada. Health Economics, 21, 1477–1495. DI MATTEO, BARBIERO Health Spending and Outcomes in Canada and Italy www.RofEA.org 435 Hitiris, T. and J. Posnett. (1992). “The determinants and effects of health expenditure in developed countries.” Journal of Health Economics 11: 173-181. Hollingsworth B.(2008) The measurement of efficiency and productivity of health care delivery. Health Economics, 17(10):1107-1128. Hurley, J. (2010) Health Economics, First Edition. McGraw-Hill Ryerson. Jacobs R, Smith PC, Street A (2006). Measuring Efficiency in Health Care: Analytic Techniques and Health Policy. Cambridge, U.K.: Cambridge University Press Joumard, I., C. André and C. Nicq (2010), “Health CareSystems: Efficiency and Institutions”, OECD Economics Department Working Papers, No. 769, OECD Publishing, Paris. Kim, D., & Saada, A. (2013). The social determinants of infant mortality and birth outcomes in Western developed nations: a cross-country systematic review. International Journal of Environmental Research and Public Health, 10(6), 2296-2335. Lavis JN. (2011). Dialogue Summary: Measuring Health System Efficiency in Canada. Hamilton, ON: McMaster Health Forum. Latif, E. (2015). Income Inequality and Health: Panel Data Evidence from Canada. The BE Journal of Economic Analysis & Policy, 15(2), 927-959. Leu, Robert E. (1986). “The public-private mix and international health care costs,” in Public and Private Health Services, edited by A.J. Culyer and B. Jonsson, 41-63. Oxford: Basil Blackwell. Lozano, R. et al., (2018) Measuring performance on the Healthcare Access and Quality Index for 195 countries and territories and selected subnational locations: a systematic analysis from the Global Burden of Disease Study 2016, The Lancet, 391, 2236-71. Lynch J, Davey Smith G, Hillemeier M, Shaw M, Raghunthan T, Kaplan G. (2001) “Income Inequality, the Psychosocial Environment, and Health: Comparisons of Wealthy Nations”. Lancet. 358:194–200. Lynch, J., Smith, G.D., Harper, S. Hillemeier, M. Ross, N., Kaplan, G.A., and Wolfson, M. (2004). Is Income Inequality a Determinant of Population Health?: A Systematic Review, The Milbank Quarterly, Vol 82, No. 1 pp.5-99 Magazinno, C. and M. Mele (2012) “The Determinants of Health Expenditures in Italian Regions” International Journal of Economics and finance, 42, March, 61-72. Marchildon, G. (2019) “Health system in Canada,” Health Services Evaluation, 769-777 https://doi.org/10.1007/978-1-4939-8715-3_41 Marchildon, G. and L. Di Matteo (2011) Health Care Cost Drivers: The Facts. Canadian Institute for Health Information. Review of Economic Analysis 12 (2020) 403-438 www.RofEA.org 436 D. Martin, A.P Miller, A. Quesnel-Vallée, N.R Caron, B. Vissandjée, G..P Marchildon (2019). “Canada’s universal health-care system: achieving its Potential,” 11-28, The Lancet. www.thelancet.com. Published online February 23, 2018 http://dx.doi.org/10.1016/S0140- 6736(18)30181-8. Medeiros, J. and Schwierz, C. (2015). The Efficiency Estimates of Health Care Systems. European Commission, Economic Paper 549. Medina, L., and Schneider, F. (2018). Shadow Economies Around the World: What Did We Learn Over the Last 20 Years? IMF Working Paper WP/18/17 Nolte, E., & McKee, M. (2004). Does healthcare save lives? Avoidable mortality revisited. London: The Nuffield Trust. Nghiem, S. H. and L.B. Connelly (2017) “Convergence and determinants of health expenditures in OECD countries”. Health Econ Rev 7, 29 (2017). https://doi.org/10.1186/s13561-017-0164-4 Nixon, J., & Ulmann, P. (2006). “The relationship between health care expenditure and health outcomes: Evidence and caveats for a causal link”. European Journal of Health Economics, 7, 7–18. Norris, S. (2018). Federal funding for Health Care. Library of Parliament. Publication No. 2018-45-E 18 July 2018. https://lop.parl.ca/staticfiles/PublicWebsite/Home/ResearchPublications/InBriefs/PDF/201 8-45-e.pdf OECD 2010, “Health care systems: Getting more value for money”, OECD Economics Department Policy Notes, No. 2. Parkin, D., A. McGuire and B. Yule. (1987). “Aggregate Health Care Expenditures and National Income: Is Health Care a Luxury Good.” Journal of Health Economics 6: 109-27. Pozen, A. and D.M. Cutler (2010) “Medical Spending Differences in the United States and Canada: The Role of Prices, Procedures and Administrative Expenses,” Inquiry, 47(2), 124- 134. Ring, I., Brown, N. (2003). The health status of indigenous peoples and others. BMJ (Clinical research ed.), 327(7412), 404–405. https://doi.org/10.1136/bmj.327.7412.404 Rowlingson, K. (2011) Does income inequality cause health and social problems?” Joseph Rowntree Foundation. www.jrf.org.uk Rubin, J. Taylor, J., Krapels, J., Sutherland, A., Felician, M., Liu, J. Davis, L., and Rohr, C. (2016), Are Better Health Outcomes Related to Social Expenditure? Rand Corporation, Cambridge, U.K. DI MATTEO, BARBIERO Health Spending and Outcomes in Canada and Italy www.RofEA.org 437 Smylie, J. and M. Firestone (2016) The health of Indigenous Peoples. In Dennis Raphael (Ed.), Social determinants of health: Canadian perspectives, 3rd edition. (434-466). Toronto, ON: Canadian Scholars’ Press. Snowden, C. (2011). The Spirit Level of Delusion: Fact-Checking the Left’s New Theory of Everything. Monday Books Stuckler D, Basu S, McKee M. Budget crises, health, and social welfare programs. BMJ. 2010;340:c3311. Available from: http://www.ncbi.nlm.nih.gov/pubmed/20576709 Subramanian, S.V. and I. Kawachi (2004) “Income Inequality and Health: What we Have Learned So Far,” Epidemiologic Reviews, 26, 1, July, 78-91. Sullivan, B. (2014) “Innovation in medical imaging can help keep health care affordable,” The Hill Times. November 17, 43. Thorne, S. (1996) “In Italy, A Medical Degree Often Means Unemployment or Underemployment.” Canadian Medical Association Journal, 154 (6), March 15, 1996. Vercelli M, Lillini R, Quaglia A, Micale RT, La Maestra S, De Flora S. (2014) Age-related mortality trends in Italy from 1901 to 2008. PLoS One. 2014;9(12):e114027. Available from: http://www.ncbi.nlm.nih.gov/pubmed/25486606 Wagner, J. & DC Aton (2012) “Incidentalomas: a "disease" of modern imaging technology.” Best Pract Res Clin Endocrinol Metab. 2012 Feb;26(1):3-8. doi: 10.1016 /j.beem.2011.08.006. Walton, J. and Binns, T.B.(ed) (1984). Medical Education and Manpower in the EEC. Fondazione Smith Kline. Wilkinson, R. and Picket, K. (2009). The Spirit Level: Why more equal societies almost always do better. Allan Lane Wilkinson, R. and Picket, K. (2009). (2015). Income Inequality and Health: A causal review. Social Science and Medicine, Vol. 128, Pages 316-326 World Health Organization (2010) Health Systems Financing: The path to universal coverage. Elizabeth H. Bradley, Maureen Canavan, Erika Rogan, Kristina Talbert-Slagle, Chima Ndumele, Lauren Taylor and Leslie A. Curry “Variation In Health Outcomes: The Role Of Spending On Social Services, Public Health, And Health Care, 2000−09,” Health Affairs 35, no.5 (2016):760-768 World Health Organization (2000). The World Health Report 2000: Health Systems: Improving Performance. Geneva: WHO Yetim,B., G. İlgün, Y. Çilhoroz, Ş. Demirci & M. Konca (2020) The socioeconomic determinants of health expenditure in OECD: An examination on panel data, International Journal of Healthcare Management, DOI: 10.1080/20479700.2020.1756112 Review of Economic Analysis 12 (2020) 403-438 www.RofEA.org 438 Appendix 1: Counterfactual Simulation Figure A1.1: Canada: Counterfactual Simulations (Log of Per Capita Total health Expenditure) Figure A1.2: Italy: Counterfactual Simulation (Log of Per Capita Total Health Expenditures) 5 6 7 8 9 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 Actual Fitted Canadian Coefficients/Italian Data 5 6 7 8 9 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 Actual Fitted Italian Coefficients/Canadian Data