In all the argument about the rights and wrongs of Julia Gillard's Fair Work Act, it's easy to forget we've been making radical changes to our industrial relations and wage-fixing system since the late 1980s.
Since the labour market is such an integral part of the economy, it's reasonable to suppose those changes have made a significant contribution to the economy's markedly improved performance over the past two decades.
What you can't do is look at the macro-economic record and confidently attribute this improvement or that deterioration to either the Fair Work Act or the Work Choices regime that preceded it.
Why not? Because, particularly if you're talking about our weak productivity performance, the timing doesn't fit. Also because neither regime was or has been in force long enough to be sure they've had much effect on anything. Changes take time to make a mark.
But, above all, because when you take just one factor and use it to explain some particular development in the economy, you're unconsciously assuming ceteris paribus - all else remains equal. And in the real world, that's never true. For all you know, the development may be explained by some other factor, or combination of factors.
So whenever you see a protagonist in the debate confidently claiming the slowdown in productivity can be blamed on the industrial relations approach they oppose, know they're talking prejudice, not reasoned analysis.
As this week's review of the Fair Work Act reminds us, we've been overhauling our industrial relations and wage-fixing system for ages. For most of the time since Federation we operated under a uniquely antipodean system of federal and state ''conciliation and arbitration''.
In theory, all strikes were illegal because the system made them unnecessary. In practice, strikes were frequent, but short.
In theory, the commission running the system would provide conciliation to differing employers and unions, trying to help them reach an agreement. If this didn't work it would compulsorily arbitrate to impose a solution, usually roughly splitting the difference. In practice, disputes were almost always settled by arbitration. Unions would pull on a strike to get the umpire to intervene and impose a decision.
In theory, wage rises were controlled by the commission in a ''national wage case'', making them nationally uniform for all workers subject to particular ''industrial awards'', on which the system was built. In practice, you could get pay rises other ways, including by breaking out of the system if you had enough muscle. Some increases would ''flow on'' from one key award to all awards.
Get the feeling it wasn't working very well? The first attempt to restore order was to limit wage rises to those awarded by the national wage case, where wages were indexed to the consumer price index.
But indexation was abandoned in 1987 and wage rises were awarded in exchange for the removal of ''restrictive work practices'' (inefficiencies). Awards were restructured to reduce demarcations between people on different awards at the same workplace and to make awards more flexible.
In 1991, the commission introduced ''enterprise bargaining'' between unions and employers at the level of the individual workplace. It applied a ''no-disadvantage [to employees] test'' before ratifying agreements. In 1994, with a new Industrial Relations Act, a more formal system of collective bargaining at the enterprise level was introduced, the national wage case was ended and replaced with a system of small annual ''safety net'' award wage increases for workers unable to negotiate an increase.
All that happened under the Hawke-Keating government. From early 1997 the Howard government's Workplace Relations Act promoted the use of individual contracts (subject to the no-disadvantage test) by introducing ''Australian workplace agreements'', increased the emphasis on enterprise bargaining by reducing the content of awards to 20 ''allowable matters'' and reduced union power by outlawing compulsory unionism and making strikes legal (''protected'') only during the negotiation of new agreements.
The Howard government's second major set of changes, Work Choices, took effect in 2006. It moved from a federal to a national system, sought to make individual contracts the main form of wage agreement by removing the no-disadvantage test, greatly increased the restrictions on unions where employers persisted with collective bargaining, and largely sidelined the commission.
After much criticism, in May 2007 John Howard significantly watered down these provisions by restoring a version of the no-disadvantage test and reinstating close scrutiny of individual contracts before approval.
Labor's Fair Work Act didn't really take effect until the start of 2010. It ended legislative recognition of individual contracts and restored collectively bargained enterprise agreements as the main form of wage-fixing. It largely restored the role of the commission under the cutesy title, Fair Work Australia. It reformed the award system, reducing more than 3000 federal and state awards to 122 simpler and less prescriptive ''modern awards''.
Some silly partisans have tried to blame Fair Work for our weak performance on the productivity of labour, but the figures don't bear this out. Productivity improved fastest under the Keating government's regime and the early years of the Workplace Relations Act, but then slowed and the weakness continued under both Work Choices and Fair Work. In any case, this ignores the huge effect of the special factors affecting productivity in mining and utilities.
But productivity isn't the only test of improved labour market performance. What about strikes? Working days lost per 1000 employees averaged 232 a year in the last days of arbitration, 176 under Labor's first Industrial Relations Act, 96 under its second act, 54 under Howard's Workplace Relations Act, 13 under Work Choices and 18 under Fair Work.
Some have blamed Fair Work for the small jump in days lost last year but most of this is explained by disputes involving the O'Farrell government and NSW public servants, who aren't covered by Fair Work.
Despite employer complaints, the minimum wage grew in real terms by less than 9 per cent over the 11 years to 2012. It dropped from 62 per cent of median full-time earnings in 1997 to 54 per cent in 2010.
The review of the Fair Work Act concludes that, since it came into force, ''important outcomes such as wages growth, industrial disputation, the responsiveness of wages to supply and demand, the rate of employment growth and the flexibility of work patterns have been favourable to Australia's continuing prosperity, as indeed they have been since the transition away from arbitration two decades ago''.