The GPT 8.125% Series A Cumulative Redeemable Preferred Stock prospectus says:
"On and after April 18th, 2012, we may redeem the Series A Preferred Stock in whole at any time or in part from time to time at a redemption price of $25.00 per share, plus any accrued and unpaid dividends to the date of redemption."
The cost would be about $135 million to redeem all the Preferred Stock and pay accrued and unpaid dividends but GPT now has $100 million in cash and the $100 million (+$50 million) Credit Facility. Redeeming the Preferred would leave GPT with $65 million to $115 million that could still be deployed to buy new properties.
The purpose would be to reduce expenses by eliminating $8 million/year in Preferred payments and pave the way for the common dividend and any equity offering. (reducing expenses adds to the bottom line)
The "Cap Rate" of redeeming the Preferreds would of course be 8.125% which is higher than the Cap Rate of the recent $20 million deal which was 7.8%+. This strategy might be deployed if highly profitable properties are harder to find.
Other options would be to redeem the Preferreds and issue new ones at a lower interest rate. GPT is in a good place. Lots of options and profitability in Q3.
Given the company's current alternatives to invest capital and it's recent credit facility, it would not be prudent to consider redeeming the preferred at this time.
Again, the reason is all about return on equity. As you suggest, redeeming the preferred generates a return on equity of 8.125%. Buying assets at a 7.8% cap rate (see my earlier post related to concerns runs about GAAP cap rates and asset pricing), the company will generate a 14% return on equity (7.8 + ((60/40) X (7.8 - 3.0)). The two alternatives are not even close.
Doing the same math, the company should begin to consider redeeming the preferred if their investment alternative is buying assets that generate an unlevered 5% return.