Ansell Crisis as Political Opportunity IU

8/13/2019 Ansell Crisis as Political Opportunity IU

1/45

"

Crisis as Political Opportunity?

Partisan Politics, Housing Cycles, and the Credit Crisis

Ben W. Ansell

University of Minnesota

The global credit crisis of 2008 was followed by a wave of successive fiscal stimulusannouncements from almost every advanced industrial state. Although the choice toengage in fiscal stimulus in response to the crisis was near-universal, the size andcomposition of these stimuli, and the accompanying panoply of financial stabilizationmeasures, varied dramatically across the OECD. This paper examines the determinants offiscal responses to the credit crisis, focusing on the interplay between government

partisanship and the relative size of the housing boom experienced by countries in theyears before the crash. I argue that volatility in asset prices like housing affects

preferences over social spending, particularly in countries with high homeownership ratesand show strong empirical evidence for these patterns on data from eighteen countriesfrom 1980 to 2003. I then build on these findings to suggest that this interactive pattern of

partisanship and housing is also evident in the varied responses of countries to the creditcrisis. In particular I show that discretionary tax cuts were largest when right-wing partiesgoverned and housing booms had been especially large, whereas discretionary spendingwas largest where left-wing parties governed and housing booms had been largest.


2/45

#

1. Introduction

The financial crisis of Fall 2008, sparking fears of a repeat of the Great

Depression, produced a dramatic and choreographed set of government responses across

the advanced industrial world. Policymakers faced immediate threats requiring urgent

action, in particular the teetering financial sector, as well as the impending likelihood of a

prolonged slump in broader economic activity. Looking to the longer run, the crisis also

galvanized debate about the regulatory framework governing both global finance, and

domestic lending. While recent economic events suggest that the worst case scenarios

were held at bay, decisions made by governments in the year following the stock-marketdownturn of October 2008 will greatly impact the long-run fiscal health, monetary

stability, and regulatory frameworks of most countries. These choices, in fact, differed

quite sharply across states. Some like the USA and China, implemented large

discretionary fiscal stimuli; others like Germany and Sweden with automatic fiscal

stabilizers kicking in, were more hesitant to provide discretionary stimuli. Still other

countries, like Greece, Iceland, Ireland, and Hungary, were forced to adopt fiscal

austerity measures in response to market fears of their insolvency. In terms of financial

stabilization there was also broad variance. Central banks and finance ministries differed

in the kinds of assets they would accept as collateral, and while some states enacted

moved to nationalize banks and regulate traders, others took a more hands off approach

to financial stabilization and regulation.

Despite the importance, and sheer magnitude, of the policies implemented by

governments since the financial crisis began, at present scholars lack both a

comprehensive cross-national account of how countries differed in their policy responses


3/45

$

and an understanding of why these different choices were made. Indeed, remarkably little

political science analysis has examined the role of governments in responding to housing

booms and busts (or, for that matter, equity market volatility). Yet, housing was the prime

cause of the credit crisis inflated house prices being the original sin of consumer over-

leverage. To properly understand how governments responded to this crisis we need to

move from the standard political economy focus on labor market income and class

differences to an analysis of how wealth and home ownership affect preferences and

policy. This paper builds on a theory of the political economy of ownership to address

two questions: (a) what explained variation in policy responses made by governments tothe credit crisis, and (b) to what degree are these responses part of a broader long-run

pattern of government responses to cycles in the housing market.

I begin in Section Two by outlining an important shift in the global economy from

a period of great price and wage volatility to one marked by a Great Moderation in the

goods and services market but with increasing volatility in the market for assets,

particularly housing and equities. Focusing on the role of housing, I theorize how

fluctuations in housing prices affect the preferences of both citizens and the political

parties that aggregate their interests and make policy. In particular, I identify a

conditional effect whereby right-wing parties are more likely to cut back taxes and

spending during periods of house price appreciation. I demonstrate the empirical strength

of this argument using cross-sectional time-series data for eighteen industrialized

countries from 1980 to 2003. I find strong and robust evidence that right-wing parties cut

social and pensions spending substantially during house price booms, even controlling for


4/45

%

other aspects of the business cycle and that this pattern is most pronounced in countries

with high levels of homeownership.

Section Three extends this analysis to cross-national responses to the credit crisis.

Does the pattern of conditional partisanship hold up in terms of the fiscal packages put

together by governments in response to the economic collapse of 2008? I begin by setting

out new data on the size and composition of the fiscal packages before turning to

empirical analysis that confirms this hypothesis. I find no direct effect of partisanship on

the nature of the stimuli but a robust conditional effect whereby countries that had

housing booms see a very strong effect of partisanship on the types of discretionary taxand spending policies they engaged in. Examining homeownership, either instead of

house price appreciation, or in combination with it, I find similar effects. Put simply,

partisan control of government mattered most in responding to the crisis where housing

formed a larger share of the economy Section Four briefly examines the financial

stabilization measures imposed by governments following the crisis, asking whether

partisanship had a direct or conditional effect on the types of regulations and support for

the financial sector adopted. Here I find only a direct effect. Section Five concludes,

discussing this distinction between fiscal and regulatory policy and offers suggestions for

our understanding of the political economy of the post-industrial but clearly not post-

financial world.

2. From Income to Assets: A Theoretical and Empirical Background

Political economists know a great deal about the political causes and

consequences of business cycles in the postwar era. Drawing first on Keynesian


5/45

&

economics and the Philips curve, and later on various versions of rational expectations

business cycle theories, successive scholars were able to draw broad conclusions both

about the effects of partisanship and labor market institutions on price inflation and

unemployment (Hibbs, 1977; Calmfors and Driffell, 1988; Hall and Franzese 1998;

Iversen, 1999), and symmetrically about the impact of these macroeconomic forces on

political outcomes from public spending to voting (Franzese, 2002; Duch and Stevenson,

2007). One irony of this expansive and convincing literature is that just as a broad

consensus on the relationship between politics and macroeconomics had emerged,

contemporary trends in the business cycle and particularly in price inflation, suggestedthat the macroeconomic volatility that characterized the mid-twentieth century had been

tamed. A variety of scholars, most notably Ben Bernanke (2004), announced a Great

Moderation had taken place with massively reduced volatility in output, employment,

and price inflation. When we examine price inflation across the OECD, that assertion

seems to hold up. Between 1956 and 1985 the average level of price inflation in the

OECD was 6.9% with a within-country standard deviation of 4.9%. Between 1985 and

2006 the average level of price inflation more than halved to 3.3%, as did the standard

deviation - to 2.4%. Put simply, price inflation has been substantially lower and

considerably less volatile over the past few decades.

However, while price inflation of goods and services was dormant, another form

of volatility was waking. From the mid-1980s, asset markets in equities and housing

became substantially more volatile. While equity markets have been volatile historically,

the same has not been true of house prices. In particular, house prices across the

advanced industrial world accelerated in hitherto unseen rates between 2001 and 2007


6/45

'

before plummeting in an unforeseen fashion between 2008 and 2010. The period between

1985 and 2006 had three times the average level of real house price inflation as that

between 1970 and 1985 and, even before 2006 and the downturn in housing prices, a

higher level of volatility. This final boom and bust period was unprecedented in modern

times. Surveying the broad sweep of American house price history, Robert Shiller (2007)

is able to locate no qualitatively comparable episodes since the surge in house prices at

the end of World War Two, which was not, in any case, accompanied by any similar

downturn.

Most analysts agree that the huge housing booms and subsequent busts across theadvanced industrial world resulted from a set of policies that enabled cheap credit,

including mortgage market and broader financial deregulation, and extended periods of

low interest rates (Schwartz, 2009). In combination these policies have produced ever-

greater leakage from the housing sector into the health of overall economy. The IMF

(2008) estimate that in a number of countries, most pronouncedly in the USA, shocks to

house prices and their feedback into consumption and residential investment account for

north of twenty percent of variation in national economic output, even though changes in

house prices do not directly figure into national income accounts. Moreover, house prices

themselves have become ever more sensitive to policy changes. In many states - not only

the Anglo-American Liberal Market Economies but also Coordinated Market Economies

like Belgium, Norway, Denmark, and Japan - over fifty percent of house price variation

can be accounted for by monetary policy changes. This phenomenon is new. The interest

rate elasticity of house prices has increased massively since the 1970s. Whereas in most

OECD states between 1970 and 1982 a one percent point increase in interest rates was


7/45

(

associated with less than a 1% decline in house prices, between 1983 and 2007 the same

interest rate change had substantially larger effects: producing a 2% decline in house

prices in Sweden, Finland and the UK, 3.5% in the USA and France, and 5% in Spain

and Holland (IMF, 2008). In short, housing was both more strongly affected by

macroeconomic policy and a greater cause of ensuing macroeconomic outcomes. The

Great Moderation was far from apparent in asset markets. It was likely this shift in

volatility from the product to the asset market that blinded policymakers to the

underlying weaknesses of the bubble economy of 2003-2008 that precipitated the credit

crisis. Hence, if we wish to understand comparative political responses to the credit crisis,it is crucial to understand conceptually how asset cycles, as opposed to regular business

cycles, affect citizens preferences and the behavior of politicians.

Elsewhere I have characterized the periods from 1945 to the early 1980s and from

the early 1980s to the present as representative of two ideal types: Employment

Dominance and Asset Dominance (Ansell, 2010b). In the former case, macroeconomic

policies have larger effects on the product market in terms of prices, wages and

employment than they do on the value of assets like equities and housing. Furthermore,

citizens in a world of Employment Dominance rely on wages for their income, and on

employer- or government-provided unemployment benefits or pensions for periods where

they are out of the labor force, either through involuntary unemployment or voluntary

retirement. Asset Dominance, by contrast, describes a world in which changes in

macroeconomic policies have a larger effect on the value of assets than on product prices,

wages, or employment. In this world, citizens depend increasingly on the value of their

assets both for day-to-day income (using the value of their asset as collateral) and for


8/45

)

unemployment or retirement (a nest egg effect). These ideal types are just that

although the evidence above is suggestive of a Great Moderation in inflation and an

accompanying growing volatility in asset prices, clearly in reality both effects exist

contemporaneously. However, the past few decades have seen a greater proximity to a

world of Asset Dominance than the postwar era.

Theoretically, how would we expect political behavior to differ in these two

environments? Iversen (2005) argues that human capital is the single greatest investment

decision engaged in by individuals. One other type of investment might provide some

competition: home purchasing. Yet whereas political economists have devotedconsiderable attention to human capital investment over the past decade (Estevez-Abe,

Iversen, and Soskice, 2000; Iversen 2005; Busemeyer, 2008; Ansell, 2010a), there has

been little discussion of the potential impact of asset ownership on political preferences.

Here I briefly lay out some expectations over the impact of asset prices on the

preferences of asset owners.

As I argue formally elsewhere (Ansell, 2007), when asset prices like housing -

are rising, asset-owning citizens will respond by desiring both lower taxes and less

publicly provided social insurance. First, increased earnings from assets are likely to

make citizens more tax-sensitive, either because their assets are directly taxed (as in

property taxation) or because they adopt the anti-tax preferences of high-income

individuals. Second, since asset owners have privately insured against income loss in

unemployment or old age by being able to use their assets especially housing as a nest

egg they have less demand for public insurance as a hedge against such risks. Private

insurance substitutes for public insurance. These responses to rising asset prices should


9/45

*

be mirrored by in the case of decreasing asset prices producing a reversal of

preferences. Arguably house price crashes should lead to increased demand for social

insurance against the vagaries of the market.

The discussion of changes in house prices begs the question of who the relevant

political actors are in this argument: after all, price changes themselves do not vote.

Homeowners are the obvious beneficiaries of house price appreciation and similarly,

renters and potential first-time buyers are the losers. These are essentially latent

political interests large groups of actors with common interests but unlikely to have

individual incentives to bear the costs of organizing for this interest. Thus, identifiableinterest groups with observable preferences over social policy are difficult to come by.

However, we might expect the aggregate national homeownership rate to affect the

balance of homeowner versus renter interests in terms of demands on politicians.

Specifically, in countries with higher homeownership rates, we would expect

homeowners to be more effective at expressing their preferences for reduced social

spending and taxation during periods of house price appreciation. Where homeowners are

a larger group they become a more attractive source of votes for politicians. Survey

analysis of the USA and the UK as opposed to Germany, finds a stronger effect of house

price appreciation on policy preferences in the former two high homeownership cases

compared to the latter country, well-known for its traditionally low homeownership rate

(Ansell, 2010b).

Who actually represents these homeowner preferences? In the empirical analysis

below I am not focusing on preferences but rather on policy outcomes, which raises the

question of how preferences are aggregated. Broadly, asset owners are represented by


10/45

"+

right-wing parties. This pattern fits the established findings that higher income and older

citizens tend to vote for, albeit not exclusively, right-wing parties. Indeed, the effects of

house price appreciation on individual preferences over social security are much stronger

for right-wing voters than for other citizens (Ansell 2010b). As such, we should expect to

see the preferences of homeowners channeled by right-wing parties during periods of

asset price change. Put simply, during house price booms I expect right-wing

governments to cut taxes and spending . In contrast, left-wing governments have fewer

asset owners among their constituents. We might expect such parties to try and increase

social insurance spending in order to target benefits to their constituents who lack the private insurance provided by asset ownership. In the empirical analysis below we find

that left-wing parties in countries with housing booms do appear to increase social

spending more than left-wing parties not experiencing housing booms.

I now turn to a brief empirical analysis of how national governments have

responded to changes in housing prices over the past few decades. 1 Even controlling for

changes in standard macroeconomic factors including national income, price inflation,

and interest rates we see powerful effects of house prices on patterns of government

spending. Not only is there a direct effect of house prices on spending but there also

appears to be a conditional effect that works through partisan control of the government.

In short, under right-wing government, the housing effect on spending is stronger. When

house prices are rising, right-wing governments appear to cut spending further and faster.

Furthermore, this pattern is most accentuated in countries with high levels of

1 A more detailed set of analyses are conducted in Ansell (2010a). The analyses below establish anempirical framework for comparison to Section Threes analysis of responses to the credit crisis.


11/45

""

homeownership. Thus, where housing is most economically important, the politics of

ownership matter substantially in social policymaking.

I examine social spending in eighteen countries between 1980 and 2001. The total

social spending measure comes from the OECD (2008) and includes pensions,

unemployment benefits, survivors benefits, incapacity benefits, health spending, family

spending, active labor market programs, and public housing. 2 The key independent

variable is the five-year percentage change in real house prices (i.e. inflation-adjusted),

taken from the Bank of International Settlements. The mean of this variable is 12.8% (i.e.

a compounded annual rate of around 2.5%), its standard deviation is 25.6% and itsminimum and maximum values are, respectively, -45.8% and 118.3%. For government

partisanship I use Cusack and Engelhardt's cabinet `center of gravity' index which

produces a measure of cabinet ideology which is the summation of the ideology of parties

comprising the cabinet weighted by their relative size in the coalition (Cusack and

Engelhardt, 2002). I use their `composite ideology' index, which is based on 23 expert

ratings. This variable ranges (theoretically) between -100 (left) and 100 (right). I also

employ an interactive variable that is the product of cabinet ideology and the five year

percentage change in house prices. My expectation is that this interaction should be

negative that is, as house prices rise under right-wing government, social spending will

be cut. In two of the models I also use homeownership data drawn from Atterhog (2005)

for fifteen of the countries.

I further include a set of macroeconomic controls including Gross Domestic

Product (measured in $100bn), the annual growth rate of GDP, the real interest rate, and

2 Ansell (2010a) examines disaggregated elements of social spending, finding that the effect of housing isstrongest on those aspects related most closely to social insurance: pensions and unemployment benefits.


12/45

"#

consumer price inflation taken from the World Development Indicators. I also add two

further important determinants of social spending: trade openness (Rodrik, 2000) and the

proportion of citizens over sixty-five years of age (Shelton, 2007). 3

Table One contains five models examining the hypothesis of an effect of

partisanship on social spending conditional on housing price appreciation. All the models

include the full array of control variables as well as year dummies (except Model A

which uses a linear time trend). Models A is a pooled analysis whereas Models B and C

include country dummies, thereby measuring only within country changes. Models A and

C both used lagged dependent variables, whereas Model B assumes an autoregressiveerror term of order one. The key finding is that the interactive variable of cabinet

partisanship and house prices is negative and robust across all three models. The

implication is that the impact of house prices, controlling for other aspects of the business

cycle and other determinants of government spending, is filtered through partisan control

of government.

To aid interpretation, Figure One shows estimates drawn from Model C of the

marginal effect on social spending of real house prices increasing by fifty percent over

five years. 4 It shows the estimated effect - with ninety-five percent confidence intervals -

of this increase in house prices for a given level of partisanship, where negative fifty

signals a government left of a countrys mean level of partisanship and positive fifty

signals a government to the right. There is a clearly negative slope: as government shifts

to the right the effect of house price appreciation on social spending moves from

3 These variables along with the dependent variables were drawn from the Quality of Governance SocialPolicy dataset compiled by Samanni et al (2008).4 Clearly using different models alters the slope and intercept of these predicted curves though broadly thesubstantive interpretation is the same.


13/45

"$

increased spending (at the extreme left) to decreased spending (for almost all

governments on the right). It is notable that this pattern is most robust for the right, with

significant decreases in spending occurring at for almost all right-wing governments.

Models D and E of Table One address the issue of homeownership. Arguably, the

impact of house price appreciation on social spending should be strongest in those

countries where homeownership is the norm, and where politicians can therefore find

more votes in addressing the policy preferences of homeowners. These models split the

sample according to whether countries have homeownership below the median (Model

D), or above it (Model E). As expected, while both sets of countries show a negativecoefficient on the interaction of partisanship and house price appreciation it is larger and

much more robust for countries with high levels of homeownership. Thus, where the

potential political coalition of homeowners is larger, fluctuations in housing prices have a

stronger effect on social policy outcomes.

Thus across the last few decades, OECD countries do appear to have experienced

important effects of housing on the behavior of governments. Does the story of the past

decades hold up today? Have right-wing parties responded similarly in the fiscal

environment following the credit crash of 2008 and have left-wing parties, for their part

moved in a countervailing direction to shore up spending?

4. Housing Booms, Partisanship, and the Stimulus

The collapse in global house prices that began in earnest in 2007 was the prelude

to a sustained decline in industrial output the following year, followed in October 2008

by a collapse in equity valuations and a near-collapse of the international financial


14/45

"%

system. Many commentators forecasted a second Great Depression. Almost

immediately, governments unveiled a set of policies to respond to these challenges. These

policies came in three forms: monetary policy easing; fiscal stimulus; and regulatory

policies. Monetary policymaking was rapid and uniform across the advanced industrial

world. By late October 2008, central banks had coordinated a massive slashing of interest

rates to near-zero levels. Policies regarding fiscal stimulus and financial regulation,

however, took longer to be enacted and varied considerably across industrial countries. In

this section and the next, I describe this variation and then examine its determinants. In

particular, I focus on the joint role of partisanship and the size of housing booms.I begin by examining the fiscal responses of advanced industrial nations to the

credit crisis. Though the crisis began in earnest in October 2008, most governments took

several months to respond fully. In particular, in February 2009 there was a rapid spurt of

announcements by governments of fiscal stimuli. The simultaneity of responses appears

related to the American Recovery and Reinvestment Act (ARRA) of the Obama

administration, enacted on Febuary 17 th 2009, having been introduced three weeks

earlier. The $787 billion act was the single largest national fiscal stimulus in absolute

amount and appears to have set the cross-national fiscal agenda. Arguably, like monetary

policy, the pattern of engaging in fiscal stimuli across the OECD displayed a high deal of

interdependence, with even the most reluctant countries (like Germany) engaging in

fiscal policy announcements in short order after the American announcement.

Thus, if we are to look for substantive cross-national variation in fiscal responses

to the credit crisis, we will not find it terms of the choice to engage in stimulus, which

was effectively universal, nor the timing, which was near-simultaneous. Instead we must


15/45

"&

to turn to examine differences among stimulus policies in terms of their size and

composition. Here we do find substantial variation. It is helpful in terms of judging the

discretionary fiscal stimuli to measure them against cyclical changes in each countrys

fiscal position caused by the credit crisis and the general economic downturn and also

against existing levels of government debt. Table Three is taken from the OECDs

analysis of the relative fiscal positions of each state (OECD, 2009).

The first thing to note is that the massive fiscal deficits carried by most states over

2009-2010 have been produced by collapses in tax revenues and countercyclical

increases in spending to cover unemployment insurance and other stabilization expenses.While the average cyclical effect on government finances was just over five percent of

GDP, discretionary stimuli averaged at only one percent of national income. However,

the relatively small average discretionary stimulus is a product of several states

(Hungary, Iceland, and Ireland) having to engage in massive fiscal retrenchment

producing an anti-stimulus effect in these states. The standard deviation for stimuli is

twice the size of that for cyclical budget deficits. Put simply, discretionary government

policies reflect substantially greater variation than the degree to which states experienced

cyclical deficits. Many states also faced further deficits from other changes which

include the purchase / nationalization of private entities by the state for example, the

nationalization of Lloyds/TSB in the United Kingdom and the nationalization of AIG in

the USA. Finally, Table Three also displays data on the level of debt held by states at the

start of the crisis in 2008 and the net addition to the national debt of 2009 and 2010.

Broadly, we see that states with higher initial debt levels have also added greater levels of


16/45

"'

debt during the downturn than those with lower debt, though some low debtors like

Ireland have had very large increases in debt.

I now turn to the composition of the fiscal stimuli. In June 2009, the OECD

released its Economic Outlook report, which provides a broad variety of information on

these fiscal packages (OECD, 2009). Table Four, using this data, details the size and

composition of the fiscal packages. Before we analyze this data in greater detail, a few

salient points jump out. Firstly, almost all countries engaged in countercyclical fiscal

stimuli in that the net change in the budget deficit was negative everywhere but Greece,

Hungary, Iceland, and Ireland, all four of which had massive capital outflows and wereforced to engage in massive budgetary retrenchment (in Greeces case not until 2010).

Elsewhere the size of stimuli ranged between less than one percent of GDP (France,

Portugal, and Switzerland) and around six percent of GDP (Korea and the USA).

There was also a great deal of variation in how the stimuli were structured both in

terms of the relative share of taxation and spending and within these categories. Some

countries tilted their stimuli towards tax cuts, for example the Czech Republic and New

Zealand (which actually cut spending), whereas others focused on spending increases, for

example Australia, Denmark, Korea, and Spain. Some countries focused tax cuts on

individual income taxes, as in Australia and Spain. Others focused on cuts to social taxes,

as in Germany and the Netherlands, or consumption taxes as in Canada and the UK.

Finally, some countries focused spending increases on government consumption

(Denmark and Sweden) others on government investment (Australia, Canada, Poland),

whereas other increased transfers to households, businesses, or subnational entities

(respectively, Luxembourg, Japan, and the United States). Clearly this is ample variation


17/45

"(

in both the size and targeting of fiscal stimuli. We now turn to how well the explanation

from the previous section related to house price booms and partisanship fits this data.

Since Cusack and Engelhardts data on cabinet partisanship does not extend until

2008, I rely instead on data drawn from Armingeon et al (2008) who code the proportion

of cabinet seats held by right-wing, center, or left-wing parties. The Armingeon data

corresponds to 2007. Thus, where control of government changed by the time of the

announced stimulus packages in OECD countries I recoded the cabinet composition

using data drawn from the collection of country profiles in Bale and Van Biezen (2009).

Below I use two different variables to reflect cabinet partisanship, along the linesdeveloped by Armingenon et al (2008). Firstly, I use right wing partisanship the

proportion of the cabinet held by right-wing politicians, following the Armingeon et al

coding of parties by country. Secondly, I use left-wing partisanship the proportion of

the cabinet held by left-wing politicians. These are not mirror images of one another

since a third category remains center-party politicians for example the Democratic

Party in the USA or the FDP in Germany. 5 I run two sets of cross-sectional regressions,

one for right-wing partisanship and one for left-wing partisanship. 6 Each time I interact

the relevant partisanship variable with a house price variable. This variable is the 2001-

2006 five-year increase in real house prices, taken from the BIS data used above. The

reason for using this span of years is that it best reflects the overall size of the housing

boom in most countries, since by 2007 prices had leveled and indeed begun declining in

5 Arguably the Democratic Party represents the left-wing in the United States. However, since the analysisin this section is cross-sectional we must be careful not to conflate the Democratic Party with socialistEuropean parties. Furthermore, the final stimulus bill in the USA reflected the necessity of overcomingsixty-vote cloture rules in the Senate, requiring two Republican votes, thereby making the bill moremoderate. Rerunning the regressions with Democrats coded as left-wing does not alter the resultssubstantially.6 Similar results are obtained by creating a partisanship index that measures right-wing cabinet proportionminus left-wing cabinet proportion.


18/45

")

many countries. Each regression thus contains partisanship, house price appreciation,

their interaction, and finally a measure of the cyclical budget deficit in 2008, drawn from

Table Three above.

I conduct two sets of regressions, one for the right-wing partisanship measure and

one for the left-wing partisanship measure. In each I examine two broad indicators of

discretionary fiscal policy first aggregate tax changes and second aggregate spending

increases. Obviously, most tax changes were negative and spending was positive, since

all these states save Ireland engaged in countercyclical stimuli. Because of Irelands

unique status as a retrencher in this cross-section, I report results both with and withoutIreland. I also show for each tax and spending dependent variable, the effects of

examining the direct effect of partisanship in Model A and then the effect conditional on

house prices in Model B. As we shall see, in every case, partisanship only matters in

terms of its interaction with housing prices.

Table Four (a) uses the right-wing partisanship variable. Models A through C

examine the determinants of discretionary tax policy. While Model A shows no net direct

effect of partisanship on the size of tax cuts, Models B and C both display an interactive

effect of partisanship conditional on the size of the housing boom in each state, albeit not

robust in Model B. Specifically, where the boom was particularly large, right-wing

government led to larger tax cuts. Or put in reverse, under right-wing control of

government states with large housing booms had large tax cuts as part of their

discretionary stimuli. The elective affinity of right-wing government, housing boom,

and cuts in revenues holds in this cross-sectional analysis. The effect is also apparent in

terms of discretionary spending. Right-wing governments in countries with large house


19/45

"*

booms had much lower spending increases than either left or center governments or,

indeed, right-wing governments in countries without large house price booms.

Table Four (b) shows symmetric results when we turn to measuring cabinet

partisanship with the proportion of cabinet seats held by left-wing parties (again

following Armingeon et als party definitions). In terms of tax stimuli the results are

statistically insignificant but the interaction term does have its expected sign positive

and as we shall see presently, there is evidence of a borderline significant impact of

housing booms when left-wing parties are entirely absent from government in this case

producing a reduction in taxes. The analysis of spending does show the expected robustinteraction between left-wing cabinet seats and house prices. When left-wing parties held

no seats house booms were associated with spending cuts in stimuli but where they held

the entire cabinet, house booms were associated with spending increases.

Since interpreting interactive coefficients is difficult, as in the previous section I

graph the estimated effects of house price booms at varying levels of cabinet

partisanship. 7 Figures Two (a) and (b), drawn from Models C and F of Table Four (a),

use the right-wing seats measure and show the estimated interactive effect of house price

booms and partisanship on, respectively, discretionary taxes and discretionary spending.

The figures show strong conditional effects. At total right-wing control of government,

house price booms are robustly associated with a decline in discretionary taxes and with a

decline in discretionary spending. Interestingly the estimated size of tax cuts and

spending cuts as a percentage of national income is almost identical implying that right-

7 The marginal change being modeled here is a doubling of house prices between 2001 and 2006. Forcomparability to Figure One, a fifty percent increase in house prices would simply be this marginal effectdivided by two.


20/45

#+

wing governments in countries with house price booms did not engage in Keynesian

countercyclical policies but accompanied tax cuts with spending cuts. 8

Figures Three (a) and (b) show the effects of left-wing control of government.

Here we see that when left-wing governments controlled cabinets there was no robust

effect of the size of the house price boom on discretionary tax policy. However at full

left-wing control of government, house price booms were associated with an increase in

discretionary spending. Put together, unlike right-wing parties, left-wing parties in

countries with housing booms did engage in counter-cyclical policies with some gusto.

However, perhaps surprisingly, this result only holds up when one examines theconditional effect of partisanship and housing prices. There is no direct effect of left-

wing (or indeed right-wing) partisanship on either discretionary tax or spending policies.

The effect of partisanship was moderated by countries housing experience.

We can view countries housing experience through another lens, that of

homeownership. Homeownership and the size of the housing boom in the first decade of

the new millennium are quite closely related. Figure Four demonstrates this pattern using

the five-year house price appreciation variable and a measure of aggregate

homeownership drawn from Atterhog (2005). 9 Arguably, parties should respond to the

wishes of homeowners more in countries where homeownership is widespread and

reflects a larger potential vote pool. Moreover, this effect should be amplified under

conditions of rising house prices as in Table One. Tables Five (a) and Five (b) use two

8 Both graphical and statistical analysis of the net change in countries fiscal position regressed on partisanship and house prices confirm this hypothesis. Left wing governments with housing booms,conversely, do appear more likely to have increased budget deficits more when housing booms were larger.9 I use data from the year available for each country in the Atterhog dataset, which varies between 1999 and2003. Hence the levels of homeownership do not correspond perfectly with those at the height of thehousing bubble. With that caveat aside, the rank order of countries in homeownership has changed verylittle over the past four decades.


21/45

#"

homeownership variables in place of the house price appreciation variable and its

interactive term used in Tables Four (a) and Four (b). 10 The first measure is the aggregate

homeownership rate and the second is that rate multiplied by house price appreciation

this latter variable measures the aggregate increase in national housing wealth.

The results are broadly similar to those found for house price appreciation, albeit

less robust with the homeownership variable and more robust with the homeownership X

house price appreciation variable. Putting these results together, it appears that aggregate

homeownership rates did indeed matter for fiscal stimulus responses but they did so

much more where house prices had also risen substantially. In terms of overalldiscretionary tax-cutting or spending increases, we find a consistent conditional effect of

partisanship, with housing shaping party responses most where ownership was broad and

price appreciation high.

Discretionary taxes and spending are still rather broad categories. It is instructive

to examine precisely what kinds of policies governments altered as part of their

discretionary stimulus policies. Tables Six (a) and (b) examine the determinants of

specific components of the stimuli. 11 Examining right-wing partisanship and taxes first

we across individual taxes, business taxes and social taxes that the general interactive

pattern we saw for taxes in general holds up it is, however, only robust in terms of

social taxation. Right wing governments in countries with house price booms appear to

have cut social insurance taxes sharply. In terms of spending we do not see any

significant effects on government consumption or investment but right-wing parties in

10 I do not split the sample as in Table One due to the very low number of cases under analysis. Thesetables should in any case be thought of as suggestive rather than inferentially rigorous.11 All the models exclude Ireland.


22/45

##

housing boom countries appear much less likely to have engaged in transfers to

households, businesses, or subnational units.

Table Six (b) repeats the exercise for the left-wing partisanship variable. Here we

see a similar pattern except in reverse to that in Table Six (a). Left wing parties in

countries without housing booms reduced social insurance taxes whereas those in

countries with housing booms did not. In terms of spending there is some evidence for

increased government consumption in states with large housing booms and left-wing

control of government. But the much stronger effects appear in terms of transfers,

especially to businesses and localities, where fully left-wing governments with house booms greatly increased transfers. Overall it appears that the interactive effects of

partisanship and house price booms appear most strongly in terms of social insurance

taxation and transfers to households.

There are two important caveats to the analysis in this section. First, whereas the

analysis in the previous section incorporated twenty years of data over eighteen countries,

we are limited in this section to a cross-section of seventeen states for which we have

data for housing prices, government partisanship and cyclical and discretionary budgetary

changes. 12 This kind of cross-sectional data forces us to make rather heroic assumptions

that partisan changes across countries are similar to those within them. However, we

obviously lack the perfect counterfactual of two different political parties in the same

country being able to simultaneously respond to the credit crisis. Thus, although the data

analysis in this section is questionable, it is at least suggestive. In fact, the pattern seen in

cross-sectional responses to the credit crisis is remarkably similar to that uncovered more

12 The BIS housing data only covers eighteen countries. The OECD does not provide data on Norwayscyclical deficit reducing the sample to seventeen. Including Norway and removing the cyclical deficitvariable does not alter results substantially.


23/45

#$

broadly in the previous section. The effect of partisanship on fiscal responses is mediated

through the relative size of the housing boom (and indeed vice versa).

Second, was the interaction between political parties and house prices in the 2009

stimuli a function of the housing boom or the bust ? If the experience of the housing boom

is more important then right-wing parties will choose fiscal responses that cut taxes

whereas left-wing parties choose those that increase spending. But potentially the parties

could have responded to the decline in house prices that began in 2008, making policy in

the reverse manner. Why did parties respond to the boom not the bust? Most stimuli were

decided upon just as house prices began to decline rather than after several years ofdecline and the empirical results in this section used the five year change in house prices,

which even in 2009 reflected much more boom than bust. And as we saw, empirical

evidence suggests that there is a hangover of house price booms right-wing parties

continued to respond in the credit crisis in the same tax-cutting, spending-trimming

fashion as during the long preceding house boom. However, it is quite possible that over

the longer run if house price declines continue we could see a reversal of preferences for

the political parties.

One way to think about this issue is to examine the cases of Japan and Germany

where house prices had in fact declined between 2001 and 2006. These cases fit closely

within the patterns found above far from being outliers, despite having very different

(in fact negative) levels of house price appreciation to the rest of the sample, Japan and

Germany across the statistical models typically have some of the lowest residuals. Thus,

the findings above are not being driven purely by house price booms those countries

with stagnating and declining house prices fit the overall pattern rather well. This of


24/45

#%

course does not directly address the question of how long it takes preferences to reverse

what is the lag between anti-tax homeowners in good times and pro-social spending

homeowners in bad times? At root, this is an empirical question that simple cross-

sectional analyses cannot address. With that said, it appears that in Japans case house

price declines did eventually force a change in preferences in line with the theory

developed above.

5. Financial Policy and Partisan Politics

I now move from fiscal policy to a brief analysis of financial policymaking during2008/2009. Perhaps the most pressing issue that governments faced in late 2008 was the

threat of a severe banking crisis turning into a complete financial meltdown. Even before

fiscal stimuli were developed, many governments had already intervened into the

financial sector through regulation and financial support. In terms of regulations,

governments quickly moved to prohibit certain forms of trading or ring-fenced particular

assets. Financial support included government purchase of toxic assets, outright

nationalization of banks, and the provision of guarantees and liquidity.

Table Seven presents descriptive statistics on these financial policies. The first

four columns examine the size of governmental support for the financial sector as a

percentage of GDP. This data is drawn from the IMF (2009) and includes direct capital

injections, the purchase of assets by national treasuries, guarantees backing financial

firms provided by governments, and the provisioning of liquidity. In all of these four

cases, governments typically did not have to draw down on the full amount promised,

largely because the financial crisis did not prove quite as deep as initially feared. The


25/45

#&

fifth column reflects the actual size of upfront financing engaged in by governments.

Whereas governments guaranteed an average of almost a quarter of national income, the

actual amount of upfront financing provided averaged only around five percent of GDP,

albeit with substantial variation. For example, whereas the Australian government only

had to provide 0.7% of GDP in upfront financing, the British government, through its

purchases of financial institutions and liquidity provision spent one fifth of national

income on financial stabilization policies.

A second way of addressing governmental intervention in the financial sector is to

simply code in a binary fashion whether the government decided to intervene in particular areas. The OECD (2009) has created one such index of financial intervention,

with the following components: (a) whether deposit insurance was enhanced; (b) whether

the government bought debt from the financial sector; (c) whether the government

engaged in direct capital injection to shore up financial firms; (d) whether the

government nationalized any financial firms; (e) whether the government ring-fenced any

assets; (e) whether the government purchased any toxic assets; (f) whether the

government decided to fund commercial paper purchases; (g) whether the government

decided to fund the purchase of securities; and (h) whether the government decided to

restrict short-selling. Clearly some of these actions are attempts to shore up financial

firms, whereas others restrict the purchase or sale of various assets. Thus, ascertaining

exactly how friendly to the financial sector such policies are as difficult some financial

firms benefit from bans on short-selling, whereas others lose. Similarly some firms

benefit from capital injections, whereas others who do not receive them may consider

them to be unfair advantages. Generally, this index is thus best interpreted as measuring


26/45

#'

the level of aggregate government intervention in the financial sector for stabilization

purposes. Again there is substantial variation across the OECD in how interventionist

governments were. The United States engaged in all of these policies whereas the Czech

Republic only banned short-selling.

To what degree were variations in financial stabilization policies a function of

partisanship, both in terms of the IMF data on the size of intervention, and the OECD

data in terms of the range of policies used? In the previous section, I found that

partisanship only explained fiscal stimulus policy conditional on a countrys housing

profile. The logic behind this argument was that home ownership and house priceappreciation change the policy preferences of citizens and that political parties

aggregating these preferences will be responding partly to the shape of a countrys

housing market. As we move to financial stabilization policies, the question is whether

we expect this pattern to continue to hold.

There are two reasons to believe that the pattern might differ. First, the fiscal

dilemmas faced by states in 2009 were closely connected to domestic housing markets

countries like Spain, the USA, and the UK that had experienced large housing booms

suddenly found a major source of employment, consumption financing, and tax revenues

drying up leading to the need for fiscal responses, whether through tax cuts (on the

right) or spending (on the left). However, the threat to financial stability from collapsing

banks affected all countries, regardless of their housing markets, since the stability of

banks largely depended on rapid flows of international capital. German banks were no

more immune from this dilemma than were British banks.


27/45

#(

Second, the preferences of political parties over financial stabilization policies are

more opaque than over taxing and spending. Financial stabilization can both harm or help

banks depending on whether they chafe at the new regulations or are bailed out by them.

Similarly, the fiscal impact of stabilization measures does not have clear consequences

for the voting public. Higher income individuals pay higher taxes and bear accordingly

greater shares of the bailout cost. However, since they are also hold a larger share of the

nations wealth in terms of savings and assets, they are also the chief beneficiaries of

explicit and implicit guarantees to the stability of the financial system. Since interest

group and voter preferences are unclear, partisan attitudes towards financial stabilizationmight instead simply reflect different party preferences over state intervention read

broadly, with left-wing parties more supportive of government regulation and ownership

of the financial market than right-wing parties. Thus our expectations for the effects of

partisanship on financial stabilization differ from those on fiscal policy. The conditional

effect of partisanship and housing markets should be considerably less important and the

direct effect of partisanship is likely to be determined not by voter preferences per se but

by party preferences over state intervention.

Accordingly, I run two sets of analyses both with and without the housing

interaction. Unlike the case of discretionary stimuli we find that the effects of

partisanship appear only to operate directly on financial stabilization without an effect

conditional on the size of the housing boom. Although some caution should be exercised

since the sample sizes across the simple bivariate and the conditional analyses are quite

distinct, it appears broadly that partisanship has a non-conditional effect on financial

stabilization policies. In particular right-wing control of the cabinet has a negative impact


28/45

#)

on the OECD composite financial regulation indicator and on both liquidity provisioning

and (less robustly) upfront government financing. Interestingly, these effects, except that

for liquidity provisioning do not hold up if we use the left-wing government partisanship

variable. In other words, right-wing governments behave differently than either left-wing

or centrist governments as regards financial stabilization but left-wing governments are

not appreciably different from centrist and right-wing governments. In fact, it is centrist

governments that are most favorable to financial regulation, although this result is mostly

driven by the USA and thus depends partly on coding the Democratic Party as centrist

rather than left-wing. Broadly, the bivariate analysis shows that right-wing governmentswere least likely to engage in either financial regulation or upfront government financing.

The multivariate analysis on the other hand displays no robust pattern of partisanship

interacted with the size of housing booms. 13 Thus, whereas the size of discretionary

stimuli appeared to depend on the combination of partisanship and housing boom size, as

regards financial stabilization partisanship simply has a direct effect. It appears that right-

wing parties unconditionally chose a less interventionist strategy than did left-wing

parties.

5. Conclusion

The credit crisis produced a flurry of activist fiscal policy across the advanced

industrial world, the likes of which had not been seen for several decades. Although, the

decision to use stimuli and their timing were near-uniform across the OECD, there was in

fact significant variation in the size and composition of fiscal stimuliIn this paper, I have

13 Very similar results are obtained using the homeownership variables in place of the house priceappreciation variables.


29/45

#*

argued that the interactive effect of government partisanship and the relative size of

housing booms experienced by these countries in the pre-crisis fat years explains a good

deal of this variation among countries. And indeed, this interactive pattern actually

explains broad patterns in social spending across the OECD over the past three decades.

Put simply, the housing market has an enormous impact of citizens wealth and on

macroeconomic outcomes more generally. It is not surprising that political parties should

internalize this effect in their policymaking. In this paper I have shown that housing

cycles, through the medium of political partisanship, appear to have strong effects on

government spending policies, particularly on social spending. The fiscal stimulus bills ofthe last few years were no exception to this general pattern.

However, in examining the determinants of financial stabilization policies I found

that the effects of partisanship were unmediated. Right-wing parties, regardless of the

size of the housing boom, were less favorable to financial regulation, and provided both

less liquidity and less upfront financing to the financial system. This might appear

surprising for two reasons. First, the stability of countries financial sectors was deeply

linked to the prevalence of cheap credit and housing booms, hence it is surprising that

house booms appear unrelated to financial stabilization measures. Second, many financial

stabilization policies could be construed as favorable to the financial sector, raising the

question of why left-wing parties engaged in policies that benefited a sector more

traditionally associated with the right. The first puzzle can be answered by reference to

the high levels of interconnectedness between the financial systems of OECD countries

when American mortgage-holders sneezed, German banks caught a cold, even though

there was no German housing boom. The second puzzle is rather more opaque since the


30/45

$+

distributive effects of financial stabilization are mixed. Perhaps left-wing parties are

increasingly supported by (or more cynically, captured by) financial sectors, as in the UK

and the USA. More likely, the greater degree of intervention by left-wing governments is

a function of broader left-wing preferences for a larger role for government, even though

the beneficiaries of financial stabilization were not typical left-wing constituents.

Bringing both fiscal and regulatory issues together, the varied responses to the

credit crisis of 2008 present an important challenge to political economists. To this point

very little theoretical or empirical work has been conducted examining the impact of

housing, or asset ownership more generally, on political behavior.14

Thus, we lack alanguage to help understand how individuals and political parties will respond to business

cycle volatility produced by the asset market, as opposed to our traditional focus, the

goods market. This paper provides a bridge linking a set of general theories about asset

markets and public spending to the specific case of the credit crisis, finding a good deal

of explanatory leverage. An equally intriguing question is how governments will react to

an era of stagnant house prices, or perhaps a long bear market, if these events come to

pass. Will the old partisan patterns of the 1960s and 1970s re-emerge, or will ownership

continue to play a key political role in the next decade?

14 For important exceptions see Schwartz (2009), Schwartz and Seabrooke (2009); Canes-Wrone and Park(2009).


31/45

$"

Bibliography Ansell, Ben, 2007, Bubbling Under: Political Preferences During Asset Bubbles,

Working Paper Ansell, Ben, 2010a, From the Ballot to the Blackboard: The Redistributive

Political Economy of Education , Cambridge University Press. Ansell, Ben, 2010b, Nest Eggs and Negative Equity: The Political Economy of

Ownership, Working Paper Armingeon, K. and Leimgruber, P. and Beyeler, M. and Menegale, S. 2009

Comparative Political Dataset I. Universitat Bern. Bale, Tim and Ingrid Van Biezen, 2009, Political Data in 2008, European

Journal of Political Research , 48:859-873. Bernanke, Ben, 2004, The Great Moderation, remarks at the meetings of the

Eastern Economic Association, Washington, February 20. Busemeyer, M. 2007. Determinants of Public Education Spending in 21 OECD

Democracies, 19802001. Journal of European Public Policy 14(2):582610.

Calmfors Lars and John Driffill, 1988, Bargaining Structure, Corporatism, and

Macroeconomic Performance Economic Policy .

Canes-Wrone, Brandice and Jee-Kwang Park, 2009, Inverse Electoral Business

Cycles and the U.S. Housing Market, Paper Presented at the Midwest Political

Science Association Annual Conference, Chicago, IL.

Clayton, Richard and Jonas Pontusson, 1997, Welfare-state retrenchment

revisited: entitlement cuts, public sector restructuring, and inegalitarian trends in

advanced capitalist societies, World Politics , 51(1):67-98.

Cusack, Thomas and Lutz Engelhardt, 2002, The PGL file collection: File

structures and procedure. Wissenschaftszentrum Berlin fr Sozialforschung

Duch, Raymond and Randolph Stevenson, 2007, The Economic Vote , Cambridge

University Press, New York, NY.

Easterly, William. 2001. Global Development Network Growth Database.

Washington DC: The World Bank. http://go.worldbank.org/ZSQKYFU6J0


32/45

$#

Estevez-Abe, Margarita, Torben Iversen, and David Soskice . 2001 . Social

Protection and the Formation of Skills: A Reinterpretation of the Welfare State.

In Peter Hall and David Soskice (eds.), Varieties of Capitalism . New York:

Oxford University Press. Franzese, Robert, 2002, Macroeconomic Policies of Developed Democracies ,

Cambridge University Press, New York, NY. Garrett, Geoffrey, 1998. Partisan Politics in the Global Economy Cambridge

University Press, New York, NY. Hall, Peter and Robert Franzese Jr., Mixed Signals: Central Bank Independence,

Coordinated Wage Bargaining and European Monetary Union, International

Organization , 2003. Hibbs, Douglas, 1977, Political Parties and Macroeconomic Policy, American

Political Science Review . IMF, 2008, The Changing Housing Cycle and the Implications for Monetary

Policy in Economic Outlook 2008. IMF 2009, Companion PaperThe State of Public Finances: Outlook and

Medium-Term Policies After the 2008 Crisis , IMF Fiscal Affairs Department. Iversen, Torben, 1999, Contested Economic Institutions , Cambridge University

Press Iversen, Torben, 2005. Capitalism, Welfare, and Democracy . Cambridge

University Press. OECD, 2009, The Effectiveness and Scope of Fiscal Stimulus, in OECD

Economic Outlook, Interim Report , June 2009. Rodrik, Dani, 2000. Why Do More Open Economies Have Larger Governments?

Journal of Political Economy 106 (5):9971032. Samanni, Marcus, Jan Teorell, Staffan Kumlin & Bo Rothstein. 2008. The QoG

Social Policy Dataset, version 24Oct08. University of Gothenburg: The Quality of

Government Institute, http://www.qog.pol.gu.se . Schwartz, Herman, 2009, Subprime Nation , Cornell University Press, Ithaca, NY.


33/45

$$

Schwartz, Herman and Leonard Seabrooke, 2008, Varieties of Residential

Capitalism in the International Political Economy: Old Welfare States and the

New Politics of Housing, Comparative European Politics . 6(3). Shelton, Cameron, 2007, The size and composition of government expenditure,

Journal of Public Economics , 91(11-12): 2230-2260. Shiller, Robert, 2007, Irrational Exuberance , Second Edition, Doubleday Press,

New York, NY.


34/45

$%

Table One: Social Spending and Housing Prices

(1) (2) (3) (4) (5)POOLED FIXED FIXED FIXED FIXED

ALL ALL ALL LOW H.O. HIGH H.O.

Lagged DV 0.950*** 0.826*** 0.734*** 0.827***(0.018) (0.040) (0.091) (0.066)

House Price Change -0.028 -0.417 -0.352 -0.582 -0.417(0.214) (0.455) (0.218) (0.470) (0.559)

Cabinet Partisanship -0.003** -0.004 -0.001 -0.004 -0.002(0.002) (0.002) (0.002) (0.004) (0.004)

Partisan*House Price -0.016*** -0.026*** -0.018** -0.014 -0.027**(0.004) (0.008) (0.007) (0.010) (0.012)

GDP per capita -0.063** -0.620*** 0.084 -0.279 0.286*(0.027) (0.189) (0.092) (0.208) (0.145)

GDP growth -0.244*** -0.051* -0.225*** -0.103* -0.282***(0.025) (0.027) (0.032) (0.059) (0.039)

Population (ln) 0.027 0.548 2.560 -13.198 -0.113(0.068) (1.998) (3.524) (11.327) (3.933)

Unemployment -0.039** 0.190*** 0.020 0.017 0.016(0.015) (0.065) (0.044) (0.083) (0.074)

Real Interest rate 0.058*** 0.059** 0.071*** 0.155** 0.083**(0.021) (0.026) (0.021) (0.061) (0.038)

Inflation -0.024 -0.049 -0.004 0.014 -0.012(0.023) (0.032) (0.029) (0.081) (0.049)

Trade -0.002 -0.032** -0.022* -0.048* -0.013(0.002) (0.016) (0.011) (0.026) (0.017)

Population>65 0.066* -0.239 -0.010 0.010 0.281(0.035) (0.290) (0.061) (0.198) (0.269)

Homeownership -0.093* -0.105(0.049) (0.071)

Constant 22.539 -4.059*** -40.063 237.965 5.423(38.094) (0.628) (58.631) (190.090) (67.180)

Observations 281 280 281 107 124R-squared 0.993 0.589 0.914 0.910 0.943

Number of countries 18 18 18 7 8


35/45

$&

Table Two: Cyclical versus Discretionary Fiscal Deficits and GovernmentDebt (all figures as % of GDP)

Cyclical

Deficit 09 Stimulus

Other

Changes

Total Deficit

Change 2008 Debt

Addition to

Debt 09-10

Australia -3.37 -3.29 -5.64 -12.30 14.24 9.99Austria -4.85 -1.21 -3.08 -9.14 65.72 10.06Belgium -7.65 -1.41 1.18 -7.88 93.18 10.38Canada -3.55 -2.61 -4.46 -10.63 68.45 10.38Czech Rep. -6.13 -2.79 2.15 -6.77 34.84 9.59Denmark -6.87 -3.28 -3.14 -13.29 40.06 6.45Finland -6.80 -3.15 -2.47 -12.42 40.61 4.15France -4.81 -0.73 -2.03 -7.56 76.05 14.36Germany -6.46 -3.19 0.48 -9.18 68.99 9.43

Hungary -7.68 7.69 -1.45 -1.45 76.47 8.17Iceland -6.40 7.27 9.39 10.26 93.21 18.32Ireland -6.27 6.48 -8.41 -8.20 47.46 22.47Italy -5.74 0.01 0.34 -5.38 114.45 10.80Japan -4.15 -4.50 -1.53 -10.18 172.09 15.52Luxembourg -7.50 -3.85 -0.81 -12.16 12.66 7.02

Netherlands -5.59 -2.54 -4.88 -13.00 64.58 11.02 New Zealand -3.91 -3.05 -6.45 -13.41 25.51 7.64Poland -4.37 -1.17 -1.20 -6.73 54.09 14.52Portugal -4.58 -0.80 -1.71 -7.09 75.26 12.46Spain -5.63 -1.34 -3.34 -10.32 46.75 17.96Sweden -8.67 -3.34 -0.58 -12.59 47.43 7.54Switzerland -3.66 -0.51 -1.72 -5.88 45.62 3.88UK -5.11 -1.26 -8.75 -15.13 57.02 26.14United States -3.07 -2.05 -4.45 -9.57 71.11 21.42

Average -5.53 -1.03 -2.19 -8.75 62.75 12.07Standard Dev 1.54 3.37 3.71 5.15 33.83 5.76


36/45

"#

Table Three: Size and Composition of Fiscal Stimuli across the OECD

Net Change TaxChanges

IndividualTax

BusinessTax

Consump.Tax

Social Tax SpendingChanges

Govt.Consump.

Govt.Investment

Transfers toHouseholds

Transfers toBusiness

Transfers toSubnational

Australia -5.4 -1.3 -1.1 -0.2 0 0 4.1 0 3 1.1 0 0Austria -1.2 -0.8 -0.8 -0.1 0 0 0.4 0 0.1 0.2 0 0.1Belgium -1.4 -0.3 0 -0.1 -0.1 0 1.1 0 0.1 0.5 0.5 0Canada -4.1 -2.4 -0.8 -0.3 -1.1 -0.1 1.7 0.1 1.3 0.3 0.1 ..Czech Rep -2.8 -2.5 0 -0.7 -0.4 -1.4 0.3 -0.1 0.2 0 0.2 0Denmark -3.3 -0.7 0 0 0 0 2.6 0.9 0.8 0.1 0 0Finland -3.2 -2.7 -1.9 0 -0.3 -0.4 0.5 0 0.3 0.1 0 0France -0.7 -0.2 -0.1 -0.1 0 0 0.6 0 0.2 0.3 0 0Germany -3.2 -1.6 -0.6 -0.3 0 -0.7 1.6 0 0.8 0.3 0.3 0Greece 0.8 0.8 0.8 0 0 0 0 -0.4 0.1 0.4 0.1 0Hungary 7.7 0.2 -0.6 -0.1 2.3 -1.5 -7.5 -3.2 0 -3.4 -0.4 -0.5Iceland 7.3 5.7 1 .. .. .. -1.6 .. .. .. .. ..Ireland 8.3 6 4.5 -0.2 0.5 1.2 -2.2 -1.8 -0.2 -0.1 0 0Italy 0 0.3 0 0 0.1 0 0.3 0.3 0 0.2 0.1 0Japan -4.7 -0.5 -0.1 -0.1 -0.1 -0.2 4.2 0.2 1.2 0.6 1.5 0.6Korea -6.1 -2.8 -1.4 -1.1 -0.2 0 3.2 0 1.2 0.7 1 0.3Luxembourg -3.9 -2.3 -1.5 -0.8 0 0 1.6 0 0.4 1 0.2 0Mexico -1.7 -0.4 0 0 -0.4 0 1.2 0.1 0.7 0.1 0 0

Netherlands -2.5 -1.6 -0.2 -0.5 -0.1 -0.8 0.9 0 0.5 0.1 0 0 New Zealand -3.7 -4.1 -4 0 0 0 -0.3 0.1 0.6 -0.6 0 0 Norway -1.2 -0.3 0 -0.3 0 0 0.9 0 0.4 0 0 0.3Poland -1.2 -0.4 0 -0.1 -0.2 0 0.8 0 1.3 0.2 0.1 0Portugal -0.8 .. .. .. .. .. .. 0 0.4 0 0.4 0Slovak Rep -1.3 -0.7 -0.5 -0.1 0 -0.1 0.7 0 0 0.1 0.6 0Spain -3.9 -1.7 -1.6 0 0 0 2.2 0.3 0.7 0.5 0.7 0Sweden -3.3 -1.7 -1.3 -0.2 0 -0.2 1.7 1.1 0.3 0.1 0 0.2Switzerland -0.5 -0.2 -0.2 0 0 0 0.3 0.3 0 0 0 0Turkey -4.4 -1.5 -0.2 -1.1 -0.2 0 2.9 0.6 1.2 0 0.3 0.6UK -1.9 -1.5 -0.5 -0.2 -0.6 0 0.4 0 0.4 0.2 0 0United States -5.6 -3.2 -2.4 -0.8 0 0 2.4 0.7 0.3 0.5 0 0.9


37/45

"#

Table Four (a): Percentage of Cabinet from Right and Stimulus Policies

MODEL A MODEL B MODEL C MODEL D MODEL E MODEL FTAX TAX TAX

EX IRE SPENDING SPENDING SPENDING

EX IRE

Right Gov 0.012 0.030** 0.021** -0.009 0.011 0.015(0.012) (0.015) (0.008) (0.009) (0.012) (0.010)

House Price Ch 0.839 0.556 0.961 1.094(0.740) (0.585) (1.200) (1.215)

Right*House Pr -0.041 -0.040** -0.046** -0.047**(0.028) (0.019) (0.019) (0.018)

Cyclical -0.301 -0.335 -0.195 0.201 0.163 0.097(0.288) (0.301) (0.185) (0.220) (0.211) (0.186)

Constant -3.266** -3.798** -2.923** 2.828* 2.226 1.814(1.493) (1.725) (1.117) (1.362) (1.395) (1.301)

Observations 17 17 16 17 17 16R-squared 0.117 0.189 0.328 0.107 0.275 0.298

Robust standard errors in parentheses *** p


38/45

"$

Table Five(a) Right Wing Partisanship and Homeownership Variables

(1) (2) (3) (4) (5) (6) (7) (8)VARIABLES Tax Tax Spend Spend Tax Tax Spend Spend

Right -0.102 -0.020 0.112* 0.086 0.023* 0.016* 0.017 0.020*(0.097) (0.107) (0.058) (0.070) (0.011) (0.007) (0.011) (0.010)

Ownership -0.000 0.017 0.041 0.036(0.038) (0.035) (0.029) (0.027)

Right*Own 0.002 0.000 -0.002* -0.001(0.002) (0.002) (0.001) (0.001)

Own*%5Yr 0.007 -0.000 0.017 0.020(0.011) (0.005) (0.015) (0.015)

Right*Own*5Yr -0.000 -0.001***

-0.001***

-0.001***

(0.000) (0.000) (0.000) (0.000)

Cyclical -0.379 -0.320 0.192 0.174 -0.309 -0.105 0.177 0.094(0.441) (0.366) (0.244) (0.246) (0.364) (0.154) (0.246) (0.211)

Constant -3.161 -3.887 -0.132 0.094 -3.568 -2.374** 2.156 1.672(3.849) (3.280) (2.360) (2.188) (2.014) (1.029) (1.624) (1.486)

Observations 16 15 16 15 14 13 14 13R-squared 0.280 0.088 0.356 0.192 0.131 0.576 0.353 0.347

Table Five (b): Left Wing Partisanship and Homeownership Variables

(1) (2) (3) (4) (5) (6) (7) (8)

VARIABLES Tax Tax Spend Spend Tax Tax Spend Spend

Left 0.152 0.017 -0.159* -0.115 -0.005 -0.007 -0.030 -0.029(0.147) (0.144) (0.088) (0.098) (0.015) (0.009) (0.021) (0.022)

Ownership 0.155 0.020 -0.131 -0.085(0.159) (0.160) (0.089) (0.103)

Left*Own -0.002 -0.000 0.002* 0.002(0.002) (0.002) (0.001) (0.001)

Own*%5Yr -0.021 -0.052*** -0.063*** -0.050**(0.043) (0.009) (0.019) (0.020)

Left*Own*5Yr 0.000 0.001** 0.001** 0.001**

(0.001) (0.000) (0.000) (0.000)Cyclical -0.563 -0.348 0.360 0.289 -0.317 -0.086 0.228 0.140

(0.464) (0.412) (0.303) (0.304) (0.355) (0.144) (0.262) (0.217)Constant -13.645 -4.660 11.256 8.262 -2.436 -1.290 4.057** 3.616**

(11.226) (11.263) (6.658) (7.426) (1.759) (0.836) (1.389) (1.263)

Observations 16 15 16 15 14 13 14 13R-squared 0.240 0.124 0.306 0.156 0.068 0.601 0.343 0.322


39/45

"%

Table Six (a): Composition of Stimuli and Right Governments

(1) (2) (3) (4) (5) (6) (7) (8) (9)IndivTax

BusTax

SocTax

ConsTax

GovCons

GovInv

TranHH

Tran Bus TranLoc

Right Gov 0.012 0.004** 0.006** -0.001 0.004 -0.001 0.002 0.009* 0.002(0.008) (0.002) (0.002) (0.002) (0.002) (0.005) (0.002) (0.004) (0.002)

House Price Ch -0.551 0.369 1.041*** -0.151 0.298 0.216 0.564 0.333 0.024(0.743) (0.241) (0.245) (0.259) (0.207) (0.694) (0.349) (0.561) (0.172)

Right*House -0.025 -0.005 -0.010*** 0.001 -0.004 -0.004 -0.015** -0.019* -0.006**(0.023) (0.003) (0.003) (0.005) (0.003) (0.008) (0.005) (0.010) (0.003)

Cyclical -0.183 -0.039 0.039* -0.036 -0.067 0.164 0.019 -0.012 0.052(0.177) (0.040) (0.021) (0.054) (0.083) (0.132) (0.046) (0.050) (0.059)

Constant -1.780 -0.640** -0.409** -0.238 -0.311 1.575 0.338 -0.031 0.385(1.065) (0.290) (0.179) (0.293) (0.523) (0.948) (0.319) (0.346) (0.429)

Observations 16 16 16 16 16 16 16 16 15R-squared 0.335 0.418 0.730 0.063 0.179 0.174 0.487 0.502 0.280

Table Six (b): Composition of Stimuli and Left Governments

(1) (2) (3) (4) (5) (6) (7) (8) (9)IndivTax

BusTax

SocTax

ConsTax

GovCons

GovInv

TranHH

Tran Bus TranLoc

Left Gov -0.005 -0.003 -0.007** 0.001 -0.009** 0.009 -0.003 -0.017** -0.009***(0.013) (0.003) (0.002) (0.003) (0.003) (0.013) (0.005) (0.007) (0.003)

House Price Ch -2.410 0.023 0.284** -0.094 -0.099 -0.089 -0.748 -1.522** -0.545***(1.555) (0.240) (0.112) (0.274) (0.281) (0.749) (0.427) (0.569) (0.115)

Left*House 0.020 0.006 0.012** -0.001 0.011* -0.004 0.015* 0.034** 0.013***(0.030) (0.005) (0.004) (0.005) (0.005) (0.017) (0.008) (0.013) (0.003)

Cyclical -0.184 -0.040 0.044* -0.037 -0.060 0.153 0.018 -0.010 0.065(0.173) (0.046) (0.023) (0.057) (0.073) (0.113) (0.054) (0.049) (0.053)

Constant -1.083 -0.386 0.025 -0.301 0.096 1.281 0.556* 0.780* 0.786*(0.820) (0.329) (0.135) (0.355) (0.390) (0.724) (0.266) (0.381) (0.361)

Observations 16 16 16 16 16 16 16 16 15R-squared 0.305 0.170 0.596 0.051 0.298 0.286 0.390 0.605 0.511


40/45

&'

Table Seven: Financial Regulations and Government Support

CapitalInjection

TreasuryAsset Purchase

Guarantees Liquidity Upfront Govt.Financing

FinancialRegulation

Australia 0 0.7 8.8 . 0.7 3Austria 5.3 0 30.1 6.4 8.9 5Belgium 4.8 0 26.4 6.4 4.8 4Canada 0 10.9 13.5 1.5 10.9 4Czech Rep . . . . . 1Denmark . . . . 5Finland . . . . 5France 1.4 1.3 16.4 6.4 1.6 4Germany 3.8 0.4 18 6.4 3.7 5Greece 2.1 3.3 6.2 6.4 5.4 3Hungary 1.1 2.4 1.1 15.7 3.5 3Iceland . . . . . 4Ireland 5.9 0 198.1 6.4 5.9 5Italy 0.7 0 0 6.4 0.7 3Japan 2.4 21.2 7.3 2.9 0.8 5Korea 2.3 5.5 14.5 4.5 0.8 3Luxembourg . . . . . 3Mexico . . . . . 1

Netherlands 3.4 10.3 33.6 6.4 13.6 5 New Zealand . . . . . 2 Norway 2 15.8 0 14.7 15.8 3Poland 0 0 3.2 5.5 0 2Portugal 2.4 0 12 6.4 2.4 5Slovak Rep . . . . . 1Spain 0 3.9 18.3 6.4 3.9 4

Sweden 2.1 4.8 47.5 13.6 5.2 4Switzerland 1.1 0 0 25.5 1.1 5Turkey 0 0.3 0 3.1 0 0UK 3.9 13.8 49.7 14.4 20 8United States 5.2 1.3 10.9 8.4 6.7 9

Mean 2.27 4.36 23.44 8.28 5.29 3.80Standard Dev 1.90 6.12 41.62 5.56 5.47 1.92


41/45

&(

Table Eight: Partisanship and Financial Responses

Bivariate

MODEL A MODEL B MODEL C MODEL D MODEL E MODEL F

FinancialRegulation CapitalInjection TreasuryPurchase Guarantees Liquidity UpfrontFinancing

Right Shareof Cabinet

-0.018** -0.013 -0.001 0.090 -0.042** -0.052*

(0.008) (0.009) (0.031) (0.202) (0.019) (0.026)Constant 4.675*** 2.845*** 4.411** 19.408*** 10.264*** 7.637***

(0.568) (0.595) (1.836) (4.771) (1.488) (1.934)


Interactive

MODEL A MODEL B MODEL C MODEL D MODEL E MODEL FFinancial

RegulationCapital

InjectionTreasuryPurchase

Guarantees Liquidity UpfrontFinancing

Right Gov -0.003 -0.021** 0.091 0.005 -0.017 -0.084*(0.010) (0.009) (0.061) (0.200) (0.069) (0.046)

House Price Ch 0.245 -2.822 2.878 5.533 0.930 0.079(1.797) (2.026) (7.390) (17.673) (7.370) (9.538)

Right*House -0.026 0.036 -0.277* 0.777 -0.051 0.082(0.022) (0.033) (0.137) (0.757) (0.155) (0.130)Constant 5.165*** 3.773*** 4.742 17.229** 10.160** 8.627**

(0.658) (0.531) (3.653) (7.229) (3.282) (3.637)



42/45

&)

Figure One: Estimated Effects of House Price Changes on Total Social Spending


43/45

&"

Figure Two (a): The Marginal Effects of House Booms on Discretionary Taxes atVarying Levels of Right Wing Government

Figure Two (b): The Marginal Effects of House Booms on Discretionary Spending atVarying Levels of Right Wing Government


44/45

&&

Figure Three (a): The Marginal Effects of House Booms on Discretionary Taxes atVarying Levels of Left Wing Government

Figure Three (b): The Marginal Effects of House Booms on Discretionary Spendingat Varying Levels of Left Wing Government


45/45

Figure Four: Homeownership Rates and House Price Booms

Documents

Ansell Crisis as Political Opportunity IU