I think this is an issue in most countries.

Government Development jobs just don't pay enough to be competitive for talent. So instead talent goes to outsourcers. Here in the Netherlands that has meant most big government development projects are done by outsourcers. Which leads to bad incentive alignment. Yielding cost overruns, projects that are DOA, and projects with way to large scope. And at the bottom line, the consultancies are better of keeping it this way than to improve things.

But government is not agile enough to pay fair wage for development work. Besides, it would be unfair to all other government employees if these developers made so much more money with less time on the job and the same level of education. (sarcasm)

Government employee in Canada here. The pay itself is not necessarily the best, but top notch benefits, generous vacation (maybe a moot point in Netherlands but definitely not here), a rock solid pension make it a lot closer than the salary number alone.

There's also the fact that the number of hours worked is always fair, and that you can move within the public service very easily.

How many big projects are handled mostly "in house" by the government in Canada. And for the outsourced ones, how much input / leadership do the internal government people have?

Because it is largely the outsourced big projects where our Government fails.

The situation is not very different here: the outsourced projects often fail, and lots of the big ones are outsourced. Definitely not a good use of public funds (my opinion), especially since the funds spent for such projects are often stratospheric. Also, the amount of talent in government is not THAT small, so often these decisions seem very political.

Exactly. Even in Malta, most government IT projects are partially outsourced to local or international companies for many of those reasons. Developers don't usually want to work for the government at the start of their career.

I think for almost all countries the taxes developers working for foreign companies pay outweight the higher costs of hiring some of them for building software for the government.

Let's say government hires 1% of the developers in the country, foreign companies pay them 3 times more and if not for outsourcing 50% of these people would have to emigrate.

If X is base salary ignoring outsourcing then the country would earn 0.5 * X * tax_rate - 0.01 * X without outsourcing and 3 * X * tax_rate - 0.01 * 3 * X with outsourcing

So the change is (3 - 0.5) * X * tax_rate - 0.01 * X * (3-1) = 2.5 * X * tax_rate - 0.02 * X = X * (2.5 * tax_rate - 0.02)

So in this simplistic example for the country to benefit from outsourcing it suffices that tax_rate > 0.008 which is pretty much always true.

Let's say government hires 10% of developers, nobody would migrate and the salary difference would be 500%. Even with such unrealistic assumptions the tax_rate would just need to be higher than 10% for the country to benefit from outsourcing.

    tax_rate > (5-1) * 0.1 / (5-1) = 0.1

If you want to play with it it's:

    tax_rate > ((salary_foreign - salary_domestic) * percent_hired_by_government) / (salary_foreign - (salary_domestic* forced_migration))

Depends how you define 'good for the country'. Upwards pressure on skilled wages due to increased demand rather than brain drain sounds like a decent place to be.