Putting up pollies’ pay?

My oped today is on politicians’ pay, reviewing the evidence on whether higher pay (a) gets better-quality people to run for office, and (b) improves the performance of legislators once they’re in office. Although the empirical findings aren’t crystal-clear, the answer to both questions seems to be  a weak yes. Full text over the fold.

Chasing Value for Money, Australian Financial Review, 16 June 2009

If Kevin Rudd were an Australian CEO, how would his pay packet compare? According to the Australian Financial Review’s Executive Salary Database, Rudd’s $330,356 would rank him 440th, just ahead of the CEO of The Reject Shop.

Should Australian politicians be paid more? At Federation, we certainly thought so. In 1901, the base salary of an Australian backbencher was £400, more than 5 times the average wage. Today, it is $127,060, less than 3 times the average wage (though MPs also enjoy other perquisites). And although our federal politicians are not yet cutting their pay (as they did in 1932), they have opted for a one-year pay freeze.

In most occupations, we have strong evidence that higher pay increases the quality of the applicant pool. But in the case of politics, there is reason to think that a bigger honeypot might attract the wrong kinds of bees. Economist Max Weber famously worried that: “Either one lives ‘for’ politics or one lives ‘off’ politics”. Perhaps a pay rise will merely increase the share of people who want to live ‘off’ politics?

Fortunately, two recent economics studies allow us to move beyond theory, and directly observe the impact of changes in politicians’ pay packets on the kinds of people who run for office, and the behaviour of legislators.

In a study of US governors, Tim Besley (London School of Economics) tested the theory that higher pay allows voters to buy representatives whose ideology is closer to their own. Analysing data across US states over three decades, Besley found that when states increase the governor’s salary packet, they are more likely to end up with a governor whose political views match those of the citizenry. Curiously, this effect seems to be partly driven by the increased probability that the governor will be a lawyer – perhaps reflecting that occupation’s ability to see the world through their client’s eyes.

New evidence from Brazil also supports the view that higher salaries raise the quality of elected officials. Cleverly exploiting a federal formula that capped the salaries of local politicians, Claudio Ferraz (PUC-Rio) and Frederico Finan (University of California, Los Angeles) found that municipalities with higher salaries attracted more candidates to run for office. More generous salaries also changed the composition of the candidate pool, boosting the average education level, the number of candidates from skilled occupations, and the share of female candidates.

Ferraz and Finan’s natural experiment also showed an effect on the productivity of legislators. Better-paid Brazilian legislators tended to submit more bills, and pass more laws. Most importantly, this did not seem to be mere busywork. Municipalities with highly-remunerated politicians had superior public services, as measured by the number of schools, health clinics and doctors.

There is also good evidence that politicians respond not merely to the level of base pay, but the generosity of retirement benefits. When the US Congress enacted a deferred increase in pensions for those who remained in office until 1992, many incumbents who would have retired in the 1990 election decided to run again, and the retirement rate leaped sharply in 1992.

In the Australian context, the old parliamentary superannuation scheme – which applied to MPs elected before 2004 – created a strong incentive for members to stay in office for at least 8 years (at which point they became eligible for a pension equal to 50 percent of salary). There is anecdotal evidence that that this stretched out the political careers of more than a few backbenchers.

Abolishing pension thresholds was probably a sensible reform, and perhaps lowered the number of timeservers warming the backbench. But the 2004 reforms also significantly cut the total remuneration available to members of parliament. Assuming the US and Brazilian results apply in Australia, the eventual effect is likely to be a lower-quality candidate pool, and perhaps even a reduction in the quality of political decisions. The same goes for the current pay freeze and mooted restrictions on electoral allowances.

Commentators sometimes make the assumption that allowing members of parliament to set their own pay is like letting the fox guard the hen-house. Yet while it is possible that politicians will seize the chance to grant themselves excessive pay increases, we should not ignore the reverse risk. When parliamentary pay comes up for public scrutiny, there is a real risk that hairshirt politics will take over, as both sides seek to outdo themselves with promises of greater austerity. Cut-price pollies may sound like a tempting reform, but it could prove costly in the long run.

Andrew Leigh is an economist in the Research School of Social Sciences at the Australian National University.

1 thought on “Putting up pollies’ pay?”

  1. While you were busy moving “beyond theory” and “empirically testing” this claim, you forgot to mention that the size of government also attracts the wrong sort of people to politics. Obviously, the more power government has, the greater is the temptation for evil people to run for office and wield that power to their own advantage.
    Whereas if you had a small state, as in the American state of New Hampshire, you could pay politicians a pittance and still attract people genuinely concerned in the public interest. They’d run for office because they cared, not to wield power.
    And, of course, you appear to be claiming that the mere passing of laws is a good measure of politicians’ productivity, given your extremely lax standard of what does and does not constitute “busywork”. That just boggles the mind. Is this really how you think? That politicians’ productivity is measurable by the number of laws they pass? Wow…*backs away slowly*


Comments are closed.