"Bad times, hard times, this is what people keep saying; but let us live well, and times shall be good. We are the times: Such as we are, such are the times."

Wednesday, 3 June 2015


Much has been made of the role of pensions reform in the last stretch of our negotiations with the troika institutions.Yet amazingly, the discussion has focused on negotiating tactics over substance - namely the viability of the Greek social security funds. In this post, partly a translation of my 22 May post on the same topic, I will attempt to shed some light on the less transparent aspects of this debate.

I have to start with the following graph, based on Eurostat data, which depicts all sources of income for Greek social security funds from 2006 to 2014. Each rectangle's area is proportionate to the amount of money it represents, providing an intuitive overview of the dependencies in our pension and benefits system. I am indebted for this idea to veteran blog reader @dalagiorgos.

The key takeaway from this graph is the enormous dependence of social security funds on subsidies, whether in the form of tax revenue ring-fenced for the funds or direct top-ups from the government budget. You see, the sum of workers' and employers' contributions into the funds and the funds' own investment income (including capital gains, interest on bonds and deposits and rents on property) came up to just 57% of the funds' income in 2014. You might dismiss this as a crisis-era aberration - after all employment is probably near all-time lows. But this share peaked at a mere 65% in the growth years (2007). Put simply, the Greek social security funds were never self-financing and social security as we know it (even in the austerity era) would cease to exist without the Government's top-up.

This life-saving subsidy to the funds amounts to EUR13bn per annum - equal to 14-15% of all Greek government revenue. A legal obligation of the state, it was, as of 2014, still up by 6% compared to the 2006-8 average, in order to compensate for a 23% drop in contributions, a 74% drop in tax revenue ring-fenced for the funds, and a 81% drop in the funds' investment income. Total flows into the funds fell by 18%.

So are the funds viable?

It might sound a little redundant to ask the question when funds are dependent on a government top-up for close to half of their income, but it is a matter of some debate whether Greek social security funds are viable. Our new government rubbished the past few years' worth of actuarial studies recently, claiming that they did not take into account the funds' non-financial portfolios. For once they are actually right; the methodology of these reports does not involve making any assumptions on returns on (non-financial) assets. It merely compares the present value of predicted in-flows and out-flows. But as I shall demonstrate, taking into account these assets would make next to no difference.

There's reasonably good data out there on the assets of social security funds, so you can test this for yourselves.

You can see the funds' financial balance sheet here. At a total of ca. EUR21bn, these assets would barely cover nine months' worth of pensions and benefits if they could somehow be liquidated under non-fire-sale conditions. Even had the funds not taken PSI losses of ca. EUR10bn in 2012 (more here), the total would still only come up to a year's worth of spending. The funds could never hope to live off the returns on this.

Estimates of the value of non-financial assets aren't as forthcoming, however a first census has been taken in 2013 as part of the implementation of the 'Hestia' database and the results can be seen here. Unfortunately, these are mark-to-algorithm values based on the Greek system of 'objective values' - which are nothing of the sort. Still, the objective values algorithm had just been rebased  to approximate market values when the Hestia census was taken in July 2013, and the Census estimate of funds' non-financial assets ran to barely EUR1.4bn. This means that the value of the funds' entire property portfolio would cover just over two weeks' worth of pensions and benefits. Nor would it be possible to close the gap merely by sweating the assets and increasing returns.

Barely 31% of the properties making up the funds' real estate empire by value were being rented out in mid-2013, and those that were yielded just 3.3% of their 'objective' values. A tenth by value were vacant, and the rest (58% by value) were reserved by the funds and wider public sector for own use. Rental income was a mere EUR20m - not one thousandth of a year's worth of fund expenditures.

As for the funds' financial assets, it's plain to see here that three quarters are made up of government bonds, cash and deposits, yielding next to nothing. This distribution is dictated by law and cost them 2-3 percentage points worth of returns per year in the good times. One sixth of the funds' financial balance sheets are made up of debtors - overdue contributions, the returns on which after defaults are presumably negative.

To cut a long story short, the total assets of the greek social security funds would not pay for ten months' worth of pensions and benefits, even if we could somehow liquidate them all without causing a fire sale. I have nothing but respect for the actuarial profession, but an actuary's work right now would not be to discuss what pensions are affordable or what the pensionable age ought to be; it would be to discuss how much of the Greek public sector's wealth would need to be injected into the funds to make them halfway viable while targeting modest outcomes.

Couldn't employers pay more? 

One of my readers objects that Greek pension funds might be more viable if employers would pay more by way of contributions. After all, at 4%-5% of GDP their contributions have consistently been slightly smaller than household contributions, and way lower than the EU average (almost half). 

I think it's a great idea to better enforce mandatory employer contributions, as a matter of principle. However, as I explain here, Greeks are much less likely to have an employer than other Europeans. We are more than twice as likely to be self-employed and three times as likely to be 'contributing family workers', some of whom are unpaid. Account for this and the fact that employer contributions in Greece are proportionately lower than in Europe becomes a matter of labour market structure as opposed to unscrupulous employers. In fact, the latest available data (2008 and 2012) suggest that social security contributions were not a low share of the total payroll cost of Greek employers by EU standards (higher than in Germany, for one) and their share of labour costs grew during the crisis. 

So increasing the headline or effective rate of employer contributions would make little financial difference, unless we could also a) increase real wages without reducing employment and/or b) restructure the Greek economy away from small and family businesses and further towards larger businesses. 

The welfare parastate

When I first posted this article, I got a very astute response from a Facebook user, which I reproduce below:

One might say that this pension top-up compensates for (typically insufficent or nonexistent) housing, unemployment and subsistence benefits for relatives (first, second or higher degree,  perhaps even friends). This social spending is instead delegated to the old and wise to distribute as they see fit among their extended family - a means, however unusual of supporting the family. 

This is 100% accurate. As I explain here, family income is more likely to be pooled in Greece than elsewhere in Europe. In the good years, the people sought, and governments gave, pension benefits in order to make up for other shortcomings of the welfare state - effectively outsourcing welfare to the pensioners. In auterity-era Greece, the heart-wrenching stereotype is of grandparents hanging their heads in shame as they find themselves unable to buy their grandkids a chocolate bar. But in reality, pension income was supporting a wider range of more basic needs, from education all the way to food.

Did Greece's welfare-delegation experiment ever work, though? This post (and the graphs to the right) summarises the outcomes using a wide range of inequality statistics for working-age and pensionable-age Greeks. In terms of narrowing or restraining inequality, the Greek welfare state only ever seemed to work for pensioners; it still did in the early crisis years.

As I keep saying, the latest data, and older data too, show that Greece’s welfare state has for years been the worst in the OECD at tackling poverty – in terms of how much it manages to reduce the risk of poverty per Euro spent. The misdirection of the pension top-up from welfare to pensions is at the heart of this weakness.

Generous or treacherous?

Famously, Greek pensions are not overly generous - at least not on a per capita basis. Greece has a lot more people of pensionable age (the WSJ takes that to mean 65+) and therefore the higher expense to match. The FT has recently reported Government claims that the average pension is EUR750 per month - which are true enough of the primary pension, but overlook the fact that one in two Greek pensioners receives more than one pension (p 6 here).

As I've explained here, Greeks on average don't retire much earlier than other Europeans, but they tend to have had shorter working lives when they do. Worse, a sizeable chunk of our labour force in major state-owned organisations have historically tended to retire very early, leaving the self-employed to lift the average retirement age.

How can we test this claim? Well, Eurostat provides detailed estimates of the share of EU countries' inactive population that are retired, by sex and age group, here. It also provides estimates of the share of the population that is inactive here. By multiplying the two, you can figure out the percentage of the population at each age group that is retired and thus entitled to some kind of pension. Run these figures and the result is astounding. Just under one in six Greeks between the ages of 50 and 59 is retired - more than 4 times the Eurozone average, and lagging only Turkey, Croatia and Slovenia. The total value of the pensions of people in their fifties comes up to almost 300m per month, and they draw the largest pensions out of any age group (p. 7 here). 

The problem isn't just that these guys (and very often ladies) will be drawing for years on benefits they can't possibly have earned. It's also that they are trapped into inactivity. Once you've retired from an office job at 50, or 55, it's fair to say you won't work again even if you wish to. 

Is this a real priority?

Our finance minister has publicly stated that reforming pensions is not a priority area for reform; our efforts, he claims would be better spent battling corruption, e.g. in public procurement. But there's a catch. The pension fund top-up (£13bn per annum as already established), is way, way larger than the entire Greek public procurement bill, right down to the last paperclip, which came up to a mere EUR8bn as of 2014. It has been larger as far back as the data go. And while not all procurement spending is unnecessary or corrupt, the pension top-up is going almost entirely to the wrong people, since it's inaccessible to, e.g., the young unemployed. The result is that Greece could be lifting more people out of poverty than it currently is through social spending, but we choose not to. Why wouldn't a left-wing government care about that? 

Where next?

Clearly, this system needs radical reform, the kind that would take years to achieve. In my view, pensions need to be cut down eventually to what the funds can pay for without further subsidy (i.e. almost cut in half). Their social policy functions should be moved instead to where they properly belong - the state budget, in the form of a means-tested guaranteed minimum income scheme. Greece piloted one such scheme, at the insistence of the Troika, in 2014. The IMF proposed this in March 2012, and Syriza - today's government - endorsed this as a matter of urgency in June 2012, before dismissing the 2014 pilot as 'crumbs of state charity', only to then reintroduce a nominal reference to a 'basic income' to its rhetoric in February 2015.

The second part of the necessary reform is to give pension funds a proper source of investment income. Syriza's Thessaloniki programme proposed to cede some of the government's real estate portfolio over to the funds. That's definitely a welcome idea. But at ca. EUR100bn on last count (2012), and even assuming every last asset can be rented out at the same nominal yield of 3.3%  as the funds' existing assets, the entire non-financial wealth of the Greek government would only give social security funds a 3.3bn lifeline per year - a fifth of what is needed to replace the current state budget top-up.

Thus, with interest rates predicted to stay low for some time, some discussion needs to be had about allowing funds to invest more of their assets in assets other than bonds and deposits. Greek stocks are volatile and pension fund allocations would move the market excessively; but overseas stocks and emerging market bonds might provide returns while also breaking the feedback loop between the state of the Greek economy and the returns on social security fund assets, which in a perfect world I'd hope would be counter-cyclical.

The third part of the reform, which I have heard not a peep about in years, would be to realign the Greek welfare state to formally acknowledge pooled (extended) family budgets and balance sheets, thus applying tax free thresholds to family as opposed to individual income, or acknowledging shared use of property and shared debt in wealth taxation. This might allow families to, for instance, set bank and utility arrears off against property tax due.

The final part is to make some provision for the people trapped into financial inactivity by early retirement. While I do not believe these are true victims of the system, they were nonetheless often lured into retirement by financial incentives, as a matter of government policy; and in any case they will not be able to earn a living again, so they are the burden of the state unless we decide to just shoot them. A means-tested guaranteed minimum income would address part of the financial problem of these individuals, but this may not be the only problem on the table.

None of this can be done quickly, or painlessly. The institutions ought to be willing to wait, and ideally to subsidise some of the transition costs, including actuarial studies, systems design and implementation. But our current predicament is such that no one trusts the Greek government with time or money unless they're bound by the conditionalities of a bailout programme. I don't trust them either.  A long-term programme, with a debt relief carrot at the end of it, just might do the trick; but things are not heading in that direction.

Either way, pensions will always be toxic as they involve a cross-subsidy from the active to the inactive, and sometimes from the poorer young to the better-off old. Within a nation state, we justify these things by appealing to people's sense of self-interest and (extended) family values and by glamorising the struggles of past generations. Across borders, none of this works.

Saturday, 23 May 2015

Whither Syriza’s Wealth Tax?

Lost amidst the tumultuous coverage of Eurogroup meetings, negotiations, leaked documents and denials, the new Greek government’s plans for tax reform are slowly taking shape and are due to be announced in the early summer of 2015. One flagship fiscal policy, announced as far back as the Thessaloniki programme, is a wealth tax. We don't know much about it, but we do know that a) it will replace ENFIA b) it will not apply to primary residences c) it is likely to be levied almost exclusively against property, although other assets might follow d) it will most likely be a net wealth tax, i.e. levied against people's equity in their belongings.

This won’t be the first attempt by Greek governments to tax wealth of course  - the hated ENFIA may have actually hastened the downfall of the previous government, prompting the kind of people that core Syriza voters had once derided as ‘νοικοκυραίοι’ to join them in the party’s ranks.

Ironically, no more than a year ago, the Bundesbank itself proposed a wealth tax (as a one-off, high impact measure) for troubled Eurozone countries. The IMF also considered a recurring wealth tax (see pg 39 here), again only to suggest that one-off levies on wealth might work better. Cyprus, of course, got a brief taste of this medicine before moving on to a more traditional depositor bail-in. 

Towards a new Wealth Tax: Taking stock

According to the ECB’s Household Consumption and Finances Survey, it took a net 331,800 in assets to put a Greek household in the top 10% of the wealth distribution (see Table A2 here) in 2009. While the threshold must have fallen very significantly since then, some of the fall in e.g. property wealth should be reversed when the economy begins to recover. So, without access to more recent figures, I will assume that it would take a threshold of EUR300,000 to capture the top 10% of Greek households by wealth. Not a bad place to start counting ‘the rich’ if you’re looking to balance left-wing legitimacy with a decent revenue, and also a number often referred to in the past as Syriza's threshold for wealth taxation.

Next is the question of how much wealth these people have. The ECB’s 2009 figures aren’t much help here, but Credit Suisse’s 2014 estimate of D10 wealth in Greece (which allows for a fat tail uncaptured by surveys) is at 56% of ca. $1 trillion, nearly half of which is in turn owned by the top 1% (pp. 125-6 here). The tax base potentially covered by our wealth tax system is therefore around EUR494 billion gross in today’s rates, or EUR462 billion net (assuming the top 10%’s ratio of gross to net wealth has not changed since 2009). 

[Note: I started writing this post in January; 'today's rates' unfortunately refer to late January]

This is the top line and it’s huge. Like an entrepreneur mulling over their business plan, the new Greek government might be tempted to think – if we can get just a little bit of that…
Well we can get a little. We know because others have tried for years. Wealth taxes are far from new, but were a dying breed in the run-up to the crisis, with more and more OECD countries abandoning them from the mid-90s onwards. Here's why.

No miracles

First, even when the tax base is wealth, taxes are paid out of income, and wealth taxes are thus constrained to some extent by the income of the wealthy. In some parts of the world (well, Germany) there is even a legal precedent of courts capping wealth tax as a share of actual income. In this context, the main function of wealth taxes is to marginally improve the fit between households’ tax liabilities and their tax-paying capacity, often by using wealth as a proxy for undeclared income. In countries like Greece, where high earners are notorious for tax avoidance, the appeal is undeniable. But large numbers of households, especially olders ones, can be asset-rich (or asset-comfortable) and cash-poor. A tragic example was one of the high-profile Greek suicides of 2012, a carer who twice insisted in his suicide note that he had ‘ample wealth, but no cash.’

Second, it's hard to pin down accurate valuations of wealth. Assets acquired in the (often remote) past may have appreciated enormously in value but, in the absence of a market sale, never been revalued accordingly. Taxpayers have no incentive to obtain independent valuations, especially under a wealth tax regime, But even if they wanted to obtain one, this will not always be possible since most of the assets that constitute wealth are highly illiquid and non-standardised. To name one example, the Spanish wealth tax system was known to consistently under-value property prices by around 70%.

Third, wealth taxes distort markets in significant ways as taxpayers seek to minimise liabilities; taxpayers can sell property and rent it back, lease yachts instead of owning outright, use equity release schemes to reduce their net equity in properties, and, of course, move their deposits (if counted as wealth) to another country.  This last option makes it hard for countries like Greece to levy any taxes on deposits unless they have the option of also implementing capital controls. I, for one, see no coincidence in the fact that Greek deposit flight accelerated following a series of news stories in January and early February claiming that Syriza was planning to tax deposits. While I cannot vouch for the quality of the sources cited by the press at the time, this plan chimed with a set of 2013 proposals made by one E. Tsakalotos, better known today as our Much More Agreeable Lead Negotiator with the Troika Institutions.

Finally, wealth taxes cost a lot to administer; they require a lot of high-quality information to run, when most states struggle to get even the basics right. Greece is struggling to get a register of property ownership together, let alone take stock of other assets.

So how much could we raise?

The tax yield from such exercises has varied across countries but as a rule it has been surprisingly small.  In Sweden, with a hundred years’ experience of wealth taxation and a vastly different tax culture to Greece’s, the wealth tax never raised more than 0.4% of GDP before being abolished. In France, it never raised more than 0.25% of GDP and in Spain it never raised more than 0.22% in recent history. The European Commission’s recent review of international wealth taxation found that revenues from net wealth taxes throughout Europe didn’t seem to exceed 0.5% of GDP in any country – about EUR1bn if applied to Greece. Historically, only small and very wealthy countries able to exercise capital controls seem to have achieved more – Switzerland probably owns the record at about 3.5% of GDP.

These examples are not entirely typical of what the Greek government could expect. Wealth taxes in other EU countries have generally focused on a narrower share of the population than the Greek government envisages and none have been implemented in the kind of struggle for survival the Greek government finds itself in. When the IMF used ECB data to calculate what might be raised on a recurring basis from a tax on the top 10%, they arrived at an estimate of around 1% of GDP across the EU (see pg 39 here). The closest equivalent to Greece might still turn out to be the case of Iceland, which reintroduced a net wealth tax in 2010, and originally set it to sunset out in 2014. We know exactly how much this raised – ca 0.9% to 1% of GDP in 2013 and 2014, despite the added support of capital controls.

That's the top-down approach. But a bottom-up estimate is also possible. My guess is that the pain threshold for such a tax is to cause taxpayers to pay as much as they would do if they had declared all income, but also not sell assets or eat into deposits, and also maintain their target ratio of liquid assets to income. If anyone is willing to try the math on this one, please leave a comment.

It’s hard to get to the liquid assets or the undeclared income of the top wealth decile in Greece, but extrapolating from the ECB’s figures for the four quartiles of the distribution back in 2009, I believe that their liquid assets might be around 3 times the group’s quarterly permanent household income (see Table 2 here for the data underlying my extrapolation). Additionally, we have estimates of unreported income by wealth quantiles from the groundbreaking work of Artavanis et al, who find that the top 10% by wealth under-report their income by ca. 50% (see Fig 1 here). That unreported income would also be part of the final tax base.

And how much would all of that that be?  Well even assuming the top 10% by wealth are also the top 10% by income (which they are definitely not) their liquid assets would not exceed 19% of Greece’s total annual household income, or ca. EUR42bn on 2013 figures, and their annual undeclared income would be around EUR28bn.

If pressed, I would say that an annual yield double that of the ENFIA income would be the best-case scenario in the medium-term, and on a very favourable set of assumptions. On balance, the risks are weighed to the downside, suggesting the wealth tax would have to be broadened, meeting fierce resistance along the way. But seriously, that's just me sticking my finger in the air. If you're willing to do the math, get in touch. 

[1] Unlike other estimates, I am tempted to use the 2009 ratios from the ECB household consumption and wealth study, because I suspect wealthier households really do target liquid assets as a share of their annual income.