This paper by a Columbia business school professor says:
The standard 60%/40% strategy outperforms a 100% bond or 100% stock strategy over the 1926-1940 period (Figure 5) and over the 1990-2011 period (Figure 6).
This is based on actual market data from those periods. You can see the figures in the PDF. These are periods of 14 and 21 years, which is perhaps shorter than the amount of time money would sit in your IRA, but still a fairly long time. The author goes on with a lot of additional discussion and claims that "under certain conditions, rebalancing will always outperform a buy-and-hold portfolio given sufficient time".
Of course, there are also many periods over which a given asset mix would underperform, so there are no guarantees here. I read your question as asking "is there any data suggesting that rebalancing a diversified portfolio can outperform an all-in-one-asset-class portfolio". There is some such data. However, if you're asking which investing strategy you should actually choose, you'd want to look at a lot of data on both sides. You're unlikely to find data that "proves" anything conclusively either way.
It should also be noted that the rebalancing advantage described here (and in your question) is not specific to bonds. For instance, in theory, rebalancing between US and international stocks could show a similar advantage over an all-US or all-non-US portfolio.
The paper contains a lot of additional discussion about rebalancing. It seems that your question is really about whether rebalancing a diverse portfolio is better than going all-in with one asset class, and this question is touched on throughout the paper. The author mentions that diversification and rebalancing strategies should be chosen not solely for their effect on mathematically-calculated returns, but for their match with your psychological makeup and tolerance for risk.