VoxEU Column Politics and economics

The wealth effect: The middle class and the changing politics of banking crises

The accumulation of mass financialised wealth has transformed the politics of banking crises. This column shows that the rising wealth of the middle classes has generated great expectations that their wealth will be protected by the government. As a result, democracies perform more financial sector bailouts and are also more financially fragile and politically unstable. 

Today, the politics of wealth is contentious in both democracies and non-democracies. It has helped to create both 'millennial socialism' and political populism (Alstadsæter et al. 2017, Ansell and Adler 2019, The Economist 2019). But wealth has been subtly reshaping politics and policy choices for decades. Our recent work shows that the accumulation of what we call 'mass financialised wealth' has made financial sector bailouts more likely and created rising financial instability – and thereby contributed to political instability and polarisation (Chwieroth and Walter 2019a, 2019b). 

It is easiest to understand these changes in the long run. Figure 1 shows how average real wealth in democracies rose sharply after WWII. Middle-class households were major beneficiaries, even in countries in which their wealth share has seen the greatest erosion recently. 

Figure 1 Real net private wealth per capita, selected countries 1870-2014, constant 2015 US$ & PPP exchange rates


Even given this plenty, other long-run developments have increased middle class anxiety (OECD 2018). 

  • More household debt. Wealth portfolios have become increasingly leveraged for segments of the middle class in advanced (and now emerging) economies. This is because the growth of mortgage credit has made it easier to buy a house, combined with rising real house prices (Jordà et al. 2016, OECD 2019: 123–124).
  • Exposure to risk. Portfolio exposure to financial risk has also risen as defined benefit pension schemes have been replaced by defined contribution schemes. 
  • Stagnating household incomes. These combine with higher relative prices for housing, education, and health care – goods associated with a middle class lifestyle with public provision after 1945 that have become increasingly marketised (OECD 2019). 

These developments have two long-run consequences for economic policy outcomes and politics that we summarise as the 'wealth effect'.

Great expectations and great disappointments

These developments have created a rising middle class stake in, and mass political demand for, financial stabilisation policies. The demand tends to be implicit during periods of relative financial tranquillity but becomes explicit during financial crises. The great expectations of wealth protection can be detected in national household surveys (Financial Services Authority 2012, van der Cruijsen et al. 2013). 

Long-run indirect evidence can be gleaned from national newspaper editorials in three countries – Brazil, the UK, and the US – since the 19th century. In all three countries, newspaper editorial writers, whether from the political right, centre, or left, shifted from supporting a relatively market-conforming policy stance during major banking crises to favouring bailouts. Figure 2 shows results for the US since 1857. Higher numbers on the vertical axis indicate stronger editorial support for government assistance to the financial system).1

Figure 2 Government policy responses and newspaper opinion in the US during ten systemic crises, 1857-2008

Source: Chwieroth and Walter (2019a, 2019b).
Notes: Newspaper document counts above bars. Newspapers: The New York Times, The Wall Street Journal, The Washington Post, The Boston Globe, The Chicago Tribune. Relevant national newspaper editorials identified using Boolean search terms that discussed government policy options during systemic banking crises since the 19th century. 

Similarly, in the long run, government policy responses in systemic banking crises have changed dramatically. Figure 3 summarises the results of our coding of thousands of government and central bank policy responses to 112 systemic banking crises in democracies since 1848. 

Figure 3 Average policy response index for new systemic banking crises and annual incidence of such crises in democracies, 1848-2010

Source: Chwieroth and Walter (2019a, 2019b).
Notes: The policy response index is the first principal component of eight policy indicators. 'Bagehot' policy responses, indicated by negative numbers in the top panel, include liquidity assistance only to solvent banks, closures of insolvent banks and losses imposed on their depositors. 'Bailout' policy responses, indicated by positive numbers, include extended liquidity provision to potentially insolvent institutions, public recapitalisations of banks, bank nationalisations, liquidity guarantees, deposit freezes, public sector bad banks, publicly subsidised relief for bank debtors, and deposit insurance schemes. Policy responses are coded in a three-year time window after the end-year of the crisis. Systemic banking crises are as defined in Reinhart (2010). Democracies are as defined by Boix et al. (2014).

It reveals a general tendency in the pre-1939 era to avoid the extensive bailouts of large financial institutions that have become the norm since the 1970s. Our more detailed statistical results support the argument that mass financialised wealth has decisively shaped this trend.

This could be consistent with the popular argument that major financial institutions have increasingly captured policymakers, and have become too big to fail (Johnson and Kwak 2010). But the influence of large banks was high in many countries before WWII, when Bagehot policies generally prevailed. Our analysis points to how the rising middle class stake in financial stabilisation has facilitated the rising influence of large financial institutions. 

In an earlier era, when mass financialised wealth was far lower, financial institutions and wealthy elites were often unable to convince elected governments to provide extensive bailouts during crises. 

Changing political consequences

If governments are now generally doing what the anxious middle class wants, have voters rewarded political incumbents? Not in practice.

The broad policy trends outlined above do not indicate that interventions are either timely or fair. Given the speed with which asset values can fall in financial crises, anxious middle-class wealth-holders with great expectations want prompt government intervention to prevent large losses at least cost. 

These expectations are often politically unrealistic. In modern democracies, extended arguments about the best policy and indecision in the early phases of crises often create large wealth losses and raise the eventual costs of bailouts. These losses help to generate the political consensus necessary to undertake financial sector interventions. 

US Congressional delays in dealing with the distressed government-sponsored enterprises (Fannie Mae and Freddie Mac) in summer 2007, and the chaos surrounding the Bush administration’s first Troubled Asset Relief Program (TARP) proposal in September 2008, generated sharp asset price declines. These losses concentrated partisan minds in Washington and so made it possible to pass TARP eventually, but neither the Bush nor the Obama administration could obtain sufficient political support for debt relief to distressed homeowners in the slow-burning mortgage crisis. The result was a bailout perceived by many as bungled, grossly unfair, and highly costly that was unable to prevent the wealth losses average Americans most feared. 

Policy delays are more likely in political systems exhibiting higher partisan polarisation and more institutional fragmentation ('veto points'). These conditions also encourage governments to intervene in ways that favour organised interests over diffuse groups such as mortgage-holders (Mian et al. 2014). 

We find that incumbent governments facing these conditions suffer higher rates of punishment after systemic banking crises only since 1970 (Figure 4).2 

Figure 4 Systemic banking crises and changes in incumbent executive party vote share, high and low veto environments, 1872 – 2011

Source: Henisz (2017), Reinhart (2010).
Notes: The results include 484 elections in 33 democracies and show the difference in the mean vote share change after systemic banking crises compared to financially tranquil times. The Henisz index captures political constraints on the chief executive based on the number of independent actors with policy vetoes and the distribution of political preferences across these actors. 

In an earlier era, free of great expectations, elected governments could survive systemic banking crises even when they failed to intervene. contrast the experience of incumbent US Republicans after deep crises in 1907 and 2007-8. In 1908, the Republican presidential successor candidate Taft and congressional allies were easily re-elected, whereas in November 2008 the Democrats swept all branches of government. 


Many reform proposals since 2008 seek to rule out future bailouts. These include tighter bank regulation, bailing-in creditors, and prohibiting or restricting government support to individual financial institutions (in the US, the Dodd-Frank Act and, in the EU, the Bank Recovery and Resolution Directive, or BRRD). 

Ben Bernanke, former chair of the Federal Reserve, and Tim Geithner and Hank Paulson, former chief secretaries of the Treasury, argue that these reforms will not prevent future crises, and that restrictions on the capacity of governments and monetary authorities to intervene are therefore dangerous (Bernanke et al. 2019). Our analysis suggests a deeper political problem: restricting crisis intervention capacity pushes against large structural trends in wealth accumulation, financialisation and electoral politics, and so lacks credibility. To illustrate, in 2017, Italy’s centre-left government set aside the brand-new BRRD. It cited a national interest loophole and avoided bailing in bondholders – one-third of whom were retail household investors – and undertook the country’s largest bank bailouts since the 1930s. 

Similar regulations are unlikely to be swept aside so easily in all cases. Instead, in highly polarised and institutionally fragmented political systems, new regulations restricting crisis interventions are likely to generate wealth-destroying policy delays in the early phase of the next crisis. Bailouts of the largest financial institutions become politically unavoidable, and if the financialised masses perceive themselves as unfairly treated, political incumbents will suffer. 

Our argument also suggests the Minskian financial cycle has deep political origins. In the long run, rising political support for bailouts has increased moral hazard. It has raised the incentives for financial institutions to insure against failure by acquiring increasingly complex and large balance sheets that are deeply connected to middle class wealth. 

Finally, our analysis helps to explain rising voter dissatisfaction with trends in public policymaking and, more broadly, in democratic processes (Foa and Mounk 2016). It suggests a new possible policy trilemma, by raising doubt about the ability of contemporary societies to sustain all three of democratic politics, rising financialisation, and financial stability over time. 

Mass wealth anxiety has been layered onto other sources of economic insecurity, notably the transformations in the labour market associated with automation and technological change, also associated with rising inequality. In an era when political consensus is fleeting, the long-run prospects for liberal democracy have become increasingly dependent on the development of broadly conceived centrist policies aimed at reducing the economic anxiety of the middle class.


Alstadsæter, A, N Johannesen, and G Zucman (2017), "Tax Evasion and Inequality", NBER working paper 23772. 

Ansell, B, and D Adler (2019), "Brexit and the Politics of Housing in Britain", The Political Quarterly 90(S2): 105–116. 

Bernanke, B S, T F Geithner, and H M Paulson Jr (2019), Firefighting: The Financial Crisis and Its Lessons, Penguin Books.

Boix, C, M K Miller, and S Rosato (2014), Boix-Miller-Rosato Dichotomous Coding of Democracy, 1800-2010. 

Chwieroth, J M, and A Walter (2019a), "The financialization of mass wealth, banking crises and politics over the long run", European Journal of International Relations. 

Chwieroth, J M, and A Walter (2019b), The Wealth Effect: How the Great Expectations of the Middle Class Have Changed the Politics of Banking Crises, Cambridge University Press.

Financial Services Authority (2012), Consumer Awareness of the FSA and Financial Regulation, Financial Services Authority.

Foa, R S, and Y Mounk (2016), "The Democratic Disconnect", Journal of Democracy 27(3): 5-17.

Henisz, W J (2017), POLCON Database.

Johnson, S, and J Kwak (2010), 13 Bankers: The Wall Street Takeover and the Next Financial Meltdown, Pantheon Books. 

Jordà, Ò, M Schularick, and A M Taylor (2016), "The Great Mortgaging: Housing Finance, Crises and Business Cycles", Economic Policy 85: 107–140. 

Mian, A, A Sufi, and F Trebbi (2014), "Resolving Debt Overhang: Political Constraints in the Aftermath of Financial Crises", American Economic Journal: Macroeconomics 6(2): 1–28. 

OECD (2018), Risks That Matter: Main Findings from the 2018 OECD Risks that Matter Survey, OECD.

OECD (2019), Under Pressure: The Squeezed Middle Class, OECD.

Reinhart, C M (2010), "This Time is Different Chartbook: Country Histories on Debt, Default, and Financial Crises", NBER working paper 15815. 

The Economist (2019), "Millennial socialism", 14 February.

van der Cruijsen, C, J de Haan, D-J Jansen, and R Mosch (2013), "Knowledge and Opinions about Banking Supervision: Evidence from a Survey of Dutch Households", Journal of Financial Stability 9(2): 219–229. 


[1] Separately, we draw on evidence from the cross-national World Values Survey since 1981 to show that there is no trend among respondents to expect governments to intervene more in other policy areas.

[2] Results are similar for coalition governments.

4,514 Reads