It seems to be accepted among utilitarians that, when dealing with charity, risk doesn't matter, as it's expected utility that matters, and the marginal utility doesn't change noticeably from your donation. This is somewhat flawed. It assumes that the only difference is what you donate, but the risks are all correlated. If there's an economic downturn, everyone donates less money, which means that there's less money going to charity, and, by the law of diminishing returns, what does go is more valuable, but you have less because of the economic downturn. If you invest in things that have risk independent of the general market that logic would work, but because of the way uncertainty adds, or rather, doesn't, that investment would be treated as low risk by the general market, so it won't be a particularly good investment.

Can anyone here translate what I just wrote into plain English?

You've lost me towards the end of your paragraph. I suppose you're saying something like:
in a downturn, the public will reduce their donation to certain charities
we should be more willing to donate to these charities because of the law of diminishing returns (the principle that it's the first few dollars that make the greatest impact)

But I don't understand how recession would affect these things. And if recession affects these things, wouldn't a boom have a counteracting effect?

I'm trying to say that if you invest your money before donating, then if you make more money, it will be needed less, and if you make less money, it will be needed more, so the expected utility is less than the utility of the expected amount of money, rather than about the same.

Think of it like this, if you're just making money for yourself, if you have a choice between definitely getting $1,000 and having a 50:50 chance between $2,000 and nothing, you should definitely go for the $1,000 because you won't get quite twice as much utility from $2,000 as you do from $1,000. People commonly say that this doesn't apply to charity because your donation is only a tiny part, so the effect is negligible. That would work if the rest of the money donated was independent of what you made, but it isn't like that. If there's a 50:50 chance of there being $1,000,000 in donations and you donate nothing, or there being $2,000,000 in donations and you donating $2,000, the $2,000 won't generate twice as much utility as $1,000 would on average, because you're only donating $2,000 if it's needed less. Hence, the effect still applies.

The argument I'm countering is generally used to say that you shouldn't try to diversify with charities, which my counter-argument doesn't really apply to, but this forum wasn't very active so I wanted to say something.

Ah, yes, I understand now.
I suppose you had asked felifician forum-goers to translate your idea into plain-english, but I think you've done it yourself, and it seems to me to be a sound argument.

For the same reason you favour non-risky investments to risky investments, you might favour counter-cyclical investments over cyclical investments.

In a recession, you will achieve a return on your investment. At the same time, other investors lose their money and so they will not be able to contribute to the revenue of charities. That means that every dollar available to be donated to charity is worth more. This is good because you have money to donate.

In a boom, you'll achieve little return on your investment, but since the revenue of charities is greater, your money would not have done so much anyway.

And, as you say, your argument doesn't give us any reason to favour a diverse portfolio of shares over a concentrated portfolio of shares, though.

Additionally, I think that in the long term (over decades), what would emerge as the most important factor is rate of return. i.e. how cyclical your portfolio is would pale in comparison.

People tend to invest in things that are going to get a good return on investment. The people who are better at it will invest more. The end result is that it's practically impossible to consistently make more money than average on an investment. The rate of return is going to be a function of risk and nothing else. Anything that's counter-cyclical makes an investment portfolio less risky, and is thus considered negative risk.

That being said, exactly how bad risk is isn't necessarily the same between people making money for themselves and charity. As such, riskier investments may be a good idea.

This general topic came up on the Old Felicifia, and I've advanced the argument that utilitarians should consider risky investments on my site. In general, I share RyanCarey's suspicion that the effect DanielLC points out is likely irrelevant in the calculations, mainly because I don't think the marginal value of charitable donations declines much even at the level of society's overall contributions. (For instance, all the utilitarians in the world donating $10 million for vegetarian outreach does still seem about twice as good as all of them donating $5 million.) Of course, this depends a lot on exactly what charitable cause is being considered.

Nevertheless, DanielLC's point is interesting. I've written about some other ways in which covariance between financial returns and other social outcomes could be relevant.