...increased government spending on high enough return investments (and there are plenty) can be maintained indefinitely even without tax increases (in rate) because the high returns over the long run allow the payoff of the government debt and much more
...high social return projects that the free market won't undertake due to market imperfections that are well established and proven in economics...like externalities, asymmetric information, impracticalities of patenting, large economies of scale and monopoly issues, the zero marginal cost of information and ideas, the inability to price discriminate well, and many more.
- The balance between investing and consuming. Investment is not a good in itself - it is valuable because it enables future consumption. Consumption now is generally regarded as better than consumption in the future - for various reasons such as the risk you won't be around later, the risk that consumption opportunities will disappear, and the simple human desire for instant gratification. So even an investment which enables lots of future consumption might not be good if present consumption needs to be suppressed to enable it. This can be modelled by maximising intertemporal utility - with consumption smoothing over a lifetime being the typical outcome.
A key point in this debate is that China already invests a lot (40-50% of GDP on some figures). So it could be that it has already found right level of investment to achieve maximum utility. The argument in fact would seem to be better applied to the developed economies (or to developing countries with a different investment profile) than to China.
- Are good investment opportunities available? Serlin seems sure that there are projects available that the government can invest in, with a positive net present value. That may well be true, but it ain't necessarily so. It is possible for all of the available profitable investment opportunities to be used up. Some would argue that this is exactly what has happened in the private sector - another debatable point, but certainly possible. So we need to show that there are still profitable places for the government to put its money. This could perhaps be done by looking at projected returns on existing discretionary government projects. If they are higher than the government's cost of capital, then we can reasonably assume that there are still opportunities out there with a return lower than existing projects, but still exceeding their financing cost.
- Does the government have the ability to pick good investments? I (and Serlin) believe it does, but the point could use some support. Principal-agency conflicts and the knowledge problem argue that the government does not make good decisions; externalities, economies of scale, and the other arguments that Serlin gives, argue that it can and should. There are market-based mechanisms available to solve most of the collective action problems that Serlin mentions, but there are also circumstances in which they are less effective than direct government action.
I haven't seen a model which attempts to quantify these arguments and compare them. It should be possible to create one, however. We could examine historical data on returns (including externalised social goods) from public investments, compared with total social return on private investments. Both are difficult to measure but I am sure some clever economist could do it. Or we could develop a theoretical model, but I'm certain that such a model could be calibrated to support either point of view, and we would need to make a judgment call about which parameters best fit the real world. At least the model would probably help us identify some success factors for government decision making, and perhaps help design democratic mechanisms to encourage good decisions.