If you are a corporation, is it risky to buy back your own debt? Most people would say no. How about if you are the Federal government? Again, it doesn't seem risky.
QE is not risky for the U.S. government (assuming the MBSs that are being purchased are already guaranteed by the Treasury). Is QE risky for the Fed? Not really, although I suppose if a person really wanted to believe it was risky, they could think of some sort of convoluted theory to explain why.
The big "risk" that people worry about is that interest rates might rise sharply, and this could cause large capital losses if the Fed's huge portfolio of bonds were sold. I see several problems with that theory:
1. First of all, the thing that would likely cause higher interest rates -- increased NGDP growth -- would actually improve the budget situation of the Federal government. Since the Fed is part of the Federal government, that's all that really matters.
2. But let's say you were really worried about the Fed's balance sheet. Some argue the Fed could simply hold the bonds to maturity. But in that case if they wanted to avoid hyperinflation they'd have to sharply increase the interest rate paid on excess reserves, which could cause losses in the short run. So holding bonds to maturity doesn't avoid losses that would be incurred by selling bonds at a loss if nominal rates rise unexpectedly. There's no free lunch there. Even so, why should we care if the Fed loses money? It's not a for-profit institution; it's part of the Federal government.
3. One fear is that the Fed might become bankrupt, as its losses on bonds might exceed its capital. It could lose its independence. However, their liabilities also include a trillion dollars of currency in circulation. Or should I say "liabilities," as it's debatable how we should regard currency. If nominal rates rise to 5%, then the Fed might have to go a few years without paying a dividend to the Federal government. But is that so bad? In the past few years they've been turning over massive profits to the Treasury, several times larger than normal. If they went a few years without paying any money to the Treasury, it would merely offset the recent excess profits. As the older low-yielding T-bonds were replaced by newer bonds yielding 5%, the Fed would turn back into the government version of Apple (AAPL) -- a money machine churning out profits of $40 or $50 billion per year.
4. Is there some tail risk I've missed? Could the losses exceed the stock of currency? Not really, but let's say they did. Suppose the U.S. shifted to electronic money just as this bear market in bonds was occurring. All currency was removed from circulation. What then? The Treasury could simply issue another trillion in T-bonds and give it to the Fed. Now this might prove embarrassing for the Fed (although I can't imagine why -- it's not their fault if the U.S. suddenly shifted to all-electronic money). But is it a problem? Clearly not. And it's not going to happen -- there'll still be a trillion in cash outstanding 10 years from now, when the Fed has adjusted to the higher interest rate.
Here's what is risky: continued tight money, which worsens the fiscal situation so much that the Treasury eventually turns the Fed into its lapdog. Tight money leading to fiscal dominance. It happened once before (1930 to 1951), let's not let it happen again. I'd add that it's very unlikely that 10 year T-bond yields will rise back to 5%. But I didn't want to rely on this argument. Trust me -- if the Fed were privatized and did an IPO, its market cap would greatly exceed Apple's.
Also, the Japanese case is far more interesting. Suppose they adopt a 2% inflation target, and their NGDP growth rises to 3%. This would be a huge boon to the Japanese government, even if the nominal interest rate on newer bonds rose by the full 3%. That's because the Japanese government, like all debtors, gains from unexpected inflation. Its existing debt was sold under expectations of no inflation. But even better for the Japanese, it's quite possible the nominal interest rate on newer debt would not rise by the full 3%, as they are currently stuck at the zero bound. Indeed rates might not rise at all. In that case, the Japanese have been walking right past 10,000 yen notes lying on the sidewalk without bothering to pick them up. I can't see that situation lasting much longer.