RWP vs RNP

Q: Why are the default intensities implied from historical data (RWP) much less than those implied from bond prices/equity prices (RNP)?

A: First, the corporate bonds are illiquid securities and bond traders demands extra returns to compensate for it. Second, bonds returns are highly skewed with limited upside. Thus, it's difficult to diversify risks in a bond portfolio than an equity portfolio. In practice, a bond portfolio is rarely fully diversified. Thus, a bond trader requires extra return for bearing this unsystematic risk in addition to the systematic risk. Third, bonds rarely default independent of each other, bonds defaults are correlated. The default rates vary from year to year depending upon economic conditions or that default by one company has a ripple effect resulting in default by other companies (credit contagion). This represents systematic risk and traders demand return for bearing this risk. 

Q: When one should use RWP and when RNP?

A: When valuing credit derivatives or  estimating the impact of default risk on the pricing of contracts one should use RNP. But for calculating expected future losses from possible defaults one should use RWP.