Put aside, for a moment, whether this is pork barrelling (hi Armidale!) or, as frontbencher Fiona Nash says, setting an example for the private sector to follow ("it shows that government thinks that community is worth investing in", she said on Wednesday).

Can it actually work?

Recent history suggests office moves will be resisted so strongly that, eventually, they either won't happen or the move will transform from an ambitious relocation to the opening of a smaller "branch office" in a regional town, with most staff staying in the city headquarters.

Look, for example, to Agriculture Minister Barnaby Joyce's portfolio, at the agencies he was determined to shift to the "bush".

Two agencies – the Murray Darling Basin Authority and the Grains Research and Development Corporation – that were to be decentralised largely remain in Canberra. The first, slated to move to Toowoomba, has now instead opened small offices in Toowoomba, Albury and Adelaide (which, with 1.3 million people, apparently qualifies as a "regional town"). Most of the authority's staff remain in the ACT. The grains corporation has also instead adopted a stay-in-Canberra model, with some new offices in Toowoomba, Dubbo, Adelaide and Perth (another "bush" town, this one with more than two million people).

Only the Rural Industries Research and Development Corporation and the Australian Pesticides and Veterinary Medicines Authority are genuine cases of decentralisation to regional areas. The corporation lost almost all its staff in when it moved to Wagga Wagga, while the pesticides authority, now in the process of shifting, is likely to suffer a similar workforce collapse.

Even the Fisheries Research and Development Corporation's relocation ("decentralised" to Adelaide) has required that it retain an office in Canberra.

Nonetheless, decentralisation can work, and has worked well at times in the past. What's crucial is that the motive for relocating is administrative rather than political.

Economist and governance expert Stephen Bartos, a former senior Finance Department official, says it requires "a logical case in terms of a well-skilled local workforce [and] suitable accommodation". He cites the example of the Tax Office, which employs about 1200 staff in its Albury offices.

"This kind of initiative works well if a department or agency itself, for business reasons, wants to locate staff in a regional area. If it doesn't, it can find good reasons not to decentralise – meaning the announced policy is likely to turn out to be more rhetoric than reality."

Another aspect of decentralisation that Bartos says the recent public debate has omitted is the involvement of state governments: will they help the transition? They can do that, he says, by providing better schools, TAFEs, university campuses (to train a high-quality workforce), transport and communications links, and social and cultural amenities.

"Otherwise, the federal government will be moving people to places where they won't be able to work effectively, meaning not only is there a cost in moving but a much larger cost in terms of reduced pubic service output. If states don't put in this investment, then the Commonwealth policy won't work."

Bartos adds that, if the relocations are forced on agencies, they will "almost certainly target lower-level employees, exacerbating divisions between senior and junior public servants – unless there is a departmental secretary out there somewhere who wants to put their hand up to move to a regional town. I'll believe this policy is more than rhetorical when that happens."

Which is pertinent for the pesticides authority, whose chief executive, Kareena Arthy, refused to tell the Senate this month whether she would move with her staff from Canberra to Armidale, but is apparently poised to jump ship. The matter was "private", she said – a response that would be fair enough for her staff to give, but which falls short of the accountability one would expect of leaders.

Finally, there's another pesky problem: whether going bush makes financial sense for agencies. Unlike businesses, government organisations have little flexibility in how much money they can pay staff, particularly under the tight controls imposed by the government's workplace bargaining framework.

As much as Canberrans might not like to admit it, many people who move here for work regard the shift as an opportunity cost – a lifestyle penalty, of sorts, given their preference to stay in their home city – for which they want to be compensated. ACT businesses put this opportunity cost at about $10,000 to $20,000 for an administrative worker (i.e. a semi-skilled, white-collar worker as opposed to a specialist). That is, Canberra employers say they need to pay about that much extra in wages to hire staff here compared with what that person would receive in a much larger city. Public service agency heads and HR managers have confirmed this to me: the quality of job applicants for, say, an APS level 6 job in Canberra is far poorer than it would be for a similar position advertised in Sydney.

So what would the labour market be like in a far smaller town than Canberra? This city is about 17 times larger than Armidale, and boasts a much wider range of amenities and opportunities. It's safe to say the opportunity cost of moving to Armidale, or a similar regional town, from a larger city would be well above $20,000. Indeed, the wages required to lure the skilled workforce that most government agencies need would be well beyond those agencies' budgets.

Unless, of course, the same government that's made a virtue of suppressing public servants' wages is suddenly willing to splash cash on bureaucrats' pay. Why not? It would be as logical as most aspects of this decentralisation push.

Markus Mannheim edits The Public Sector Informant.