Park your politics at the door for this one. Yes, I know that will be hard for some. But it's OFF TOPIC.

[Update: now that the macroeconomics have been thoroughly discussed in comments, I can relax this. Civilised discussion of politics is now OK.]

Yes, I understand Keynesian macroeconomics at least as well as most Keynesians. In particular, I understand the Keynesian case against general wage cuts at least as well as most Keynesians. I understand why cuts in nominal wages are generally not a solution to unemployment caused by a deficiency of aggregate demand. They don't work.

But there seems to me to be one special case where wage cuts could increase employment, even if unemployment is due solely to deficient aggregate demand, even in a purely Keynesian model.

The argument is embarrassingly simple. I'm almost ashamed to put it forward. There seems to be almost no economics in it. It seems to be just arithmetic. Cut government wages by 1% and increase government employment by 1% at the same time. Don't change anything else.

[Update: check out very good comments by david and rsj below. To paraphrase: david says "But there's presumably a micro-reason for wage stickiness. Private firms won't cut wages by 1% because they know that (for some reason) productivity will fall by 1% or more if they do. If this same micro-reason also applies to government workers, this policy will not increase output."; and rsj says "It's equivalent to the balanced budget multiplier. It's like a tax on government workers and the proceeds being used to increase govenrment expenditure."]

Hold government spending constant. That doesn't mean government spending will be constant over time. It means "do not change government spending on account of this policy". If you were going to increase government spending by 10%, then go ahead and do it anyway.

Hold taxes and transfers constant. That doesn't mean taxes and transfers will be constant over time. It means "do not change taxes and transfers on account of this policy". If you were going to cut taxes by 10%, and increase transfers by 10%, then go ahead and do it anyway.

Hold monetary policy constant. That doesn't mean monetary policy will be constant over time. It means "do not change monetary policy on account of this policy". If you were going to cut interest rates by 3%, then go ahead and do it anyway.

Now, cut wages of government workers by 1%, and hire 1% more government workers. Again, that is relative to what you would have done anyway.

You have just increased employment, and reduced unemployment.

The benefits are increased government output. If government workers have positive value marginal product — if extra workers produce extra outut, and it's useful output — that's a good thing. (A minor additional benefit is that it would improve the distribution of income, if you make the reasonable assumption that employed government workers are better off than the unemployed.)

I can't think of any plausible reason why it wouldn't work.

By assumption, there is demand-deficient unemployment, so there should be no difficulty in general of hiring 1% more government workers.

By construction, there are zero budgetary consequences. (Sure, the cost of a worker is pay plus benefits, so you might want to read my "cut wages by 1%" as "cut pay plus benefits by 1%".) No change in government spending, and no change in taxes and transfers. In the simplest Keynesian model nothing else will change. There's a few more people employed, usefully producing government output of goods and services, and a few less unemployed, but otherwise the private sector should be unaffected.

Sure, the existing government workers will have lower income and will cut their demand for consumption goods, but the newly hired workers will have higher income and will increase their demand for consumption goods, so it should be a wash. I can't think of any good reason why the former should have a higher marginal propensity to consume than the latter. And remember, I am assuming taxes and transfer payments stay constant. The newly-hired government workers will lose their unemployment benefits, but the government can use those savings to increase transfer payments somewhere else.

If the increased government output were a close substitute for private consumption or investment spending, then it is conceivable that private demand for output and labour would fall dollar for dollar with increased government output and employment. But no Keynesian would be caught dead making that version of the crowding out argument. The goods and services the government provides are useful, but generally very different from the goods and services we buy privately. Plus, it's a free lunch. Would Keynesians turn down a free lunch? The whole point of Keynesian macroeconomics is to make sure the free lunches available in a recession get eaten. Plus, the more free lunches we get, the wealthier we will feel, and maybe we will spend more on other stuff, which would increase aggregate demand for private sector output and employment?

Why are government wage cuts different from private sector wage cuts? Because government spending and employment is exogenous, while private spending and employment is endogenous. In a Keynesian recession, private sector wage cuts will only increase employment if they increase the demand for private sector output. And, in general, they won't do that. But in a Keynesian model, government sector output is an exogenous policy variable. So the thought-experiment of cutting government wages by 1% and increasing government employment by 1% is an internally consistent one.

Some might try a public choice/political economy response. "But if the government cuts wages by 1% it won't spend the money to hire more workers". OK, but if you believe that government employment is endogenous, then why are you wasting your breath telling the government to spend more? And, if government employment is endogenous, mightn't the government have a downward-sloping demand curve for labour?

You might argue that there's a better way to cure demand-deficient unemployment. And I would agree (though I would generally prefer using monetary rather than fiscal policy). But if the government isn't increasing expenditure and/or cutting taxes as much as you want it to, because it is scared of the deficit (even if you think that is an unreasonable fear) wouldn't you, as a good Keynesian, nevertheless support cuts in government wages and equal increases in government employment as a second best? Because you could convince the recalcitrant government that this Keynesian policy would not make the deficit worse.

Again, please keep to economics in the comments. I am not interested in 101 opinions on Stephen Harper(?) or Scott Walker(?). Does this policy make macroeconomic sense?