August 23, 2007
Via EDGAR (Correspondence)
Ms. Joyce Sweeney
Division of Corporation Finance
United States Securities and Exchange Commission
100 F Street, N.E.
Washington, D.C. 20549
Re: State Street Corporation
Form 10-K for the Fiscal Year Ended December 31, 2006, filed on February 20, 2007
Form 10-Q for the Fiscal Quarter Ended March 31, 2007, filed on May 4, 2007
File No. 001-07511
Dear Ms. Sweeney:
This letter sets forth the response of State Street Corporation (“State Street” or “we”) to the comments of the Staff of the Division of Corporation Finance of the Securities and Exchange Commission (the “Staff”) as set forth in the Staff’s letter of July 26, 2007 (the “Comment Letter”) regarding the above-referenced periodic reports. The Comment Letter relates to State Street’s response dated July 20, 2007 to the Staff’s letter of June 21, 2007 regarding the above-referenced reports.
For reference purposes, the text of each of the Staff’s numbered comments in the Comment Letter has been provided herein in bold. Our responses follow each of the comments.
Form 10-K for the Fiscal Year Ended December 31, 2006:
Consolidated Financial Statements
Note 15 — Derivative Financial Instruments, page 95
1. We note your response to comment one of our letter dated June 21, 2007. Please quantify for us the changes in fair value related to fair value hedges, the changes in fair value of the hedged items, and the amount of ineffectiveness related to cash flow hedges for each period presented.
The table below presents the amounts related to changes in fair value associated with fair value hedges and changes in fair value of the corresponding hedged items. The table separates the impact of the hedge strategies employed by our Global Markets division and by our Global Treasury group, as we discussed in our response dated July 20, 2007 to Staff comment numbers 4 and 6 in the Staff’s letter of June 21, 2007.
There were no amounts recorded for ineffectiveness associated with cash flow hedges for any period presented because, as we noted in our response dated July 20, 2007 to Staff comment number 4 in the Staff’s letter of June 21, 2007, we applied the “shortcut” method to all cash flow hedges employed during those periods. These cash flow hedges were all employed by State Street’s Global Treasury group; State Street’s Global Markets division has not employed any cash flow hedge strategies.
DERIVATIVE FINANCIAL INSTRUMENTS -
IMPACT OF HEDGE ACCOUNTING
Years ended December 31, |
| 2006 |
| 2005 |
| 2004 |
| |||
(in thousands) |
|
|
|
|
|
|
| |||
|
|
|
|
|
|
|
| |||
Global Markets division |
|
|
|
|
|
|
| |||
Changes in fair value associated with fair value hedges |
| $ | 42,785 |
| $ | 38,743 |
| $ | 4,692 |
|
Changes in fair value of hedged items |
| (42,027 | ) | (36,080 | ) | (4,488 | ) | |||
|
|
|
|
|
|
|
| |||
Global Treasury group |
|
|
|
|
|
|
| |||
Changes in fair value associated with fair value hedges |
| 8,696 |
| (32,195 | ) | 2,850 |
| |||
Changes in fair value of hedged items |
| (9,139 | ) | 32,149 |
| (2,819 | ) | |||
As we noted in our response dated July 20, 2007 to Staff comment number 1 in the Staff’s letter of June 21, 2007, amounts related to changes in fair value associated with fair value hedges, and changes in fair value of the corresponding hedged items, associated with qualifying hedge derivatives are recorded in processing fees and other revenue in our statement of income. Processing fees and other revenue for the years ended December 31, 2006, 2005 and 2004 were $272 million, $302 million and $308 million, respectively.
2. We note your response to comment four in our letter dated June 21, 2007. In your description of fair value hedges of recognized fixed rate liabilities you indicate that you performed an initial assessment of hedge effectiveness. Please describe for us the specific quantitative measures you used at inception to perform your initial assessment. If you used a statistical method, tell us the specific statistical outputs you considered.
The one strategy for which we did not apply the “shortcut” method was actually eligible for the “shortcut” method and met all the criteria of paragraph 68 of SFAS No. 133. We applied a “long haul” approach instead and recorded ineffectiveness associated with this strategy in our statement of income. This ineffectiveness was not material, and is reflected in the amounts presented in the table included in our response to Staff comment number 1 above. We did not choose to apply a statistical method at inception as permitted by Statement 133 Implementation Issue No. E7 (“Issue E7”), but rather, applied a dollar-offset approach to both our assessment of hedge effectiveness and our measurement of hedge ineffectiveness, as also permitted by Issue E7.
Accordingly, with respect to this strategy, we acknowledge that we must abide by the result of the dollar-offset calculation and therefore, conceivably could have lost hedge accounting for any period in which the dollar-offset calculation was not within an acceptable range (we are using 80% to 125%), even if we fell outside the acceptable range due to the “law of small numbers” phenomenon. Thus far, all of our dollar-offset calculations have fallen within the acceptable range because our relationships contained all the matching terms characteristic of the “shortcut” method, and because thus far interest rate movements have been significant enough each period such that the small numbers phenomenon has not dominated our dollar-offset results.
2
3. As a related matter, you indicate that for the relationships where the “shortcut” method was not applied you performed monthly effectiveness assessments using the cumulative dollar-offset approach. Please tell us whether there were any times you failed the dollar offset method and yet still concluded that the relationship is highly effective and continue to apply hedge accounting. If so, tell us what additional procedures you performed and how you concluded it was appropriate to continue to apply hedge accounting.
There were no times during the reported periods in which we failed the dollar-offset method with respect to hedge strategies employed by our Global Treasury group. Our results for our Global Treasury group in all periods between 2004 through 2006 ranged from 87.8% to 109.7%, within our acceptable range of 80% to 125%.
There were times during the reported periods at which we failed the dollar-offset method with respect to certain effective duration portfolios related to the municipal security hedge strategy employed by our Global Markets division. In these situations, we applied a dollar tolerance threshold to address the “law of small numbers” phenomenon. We deemed it appropriate to continue to apply hedge accounting in future periods since all of the effective duration portfolios were re-established monthly with a new instrument composition and designation.
The fair value adjustments to the hedged securities were so small in actual dollar amount that there was no material difference between applying fair value hedge accounting and not applying fair value hedge accounting. Specifically, in 2004 through 2006, we noted that these actual fair value changes each month ranged from $(850,557) to $411,969, none of which were material compared to quarterly processing fees and other revenue that ranged from $61 million to $84 million during the same periods. Systematically, it would have been difficult to stop and start fair value hedge accounting just because we failed the dollar-offset method for such immaterial movements in fair value. Our similar effective duration hedge matching strategy (explained in more detail in our response below to Staff comment number 4), which reset every fair value hedge every month, permitted us to maintain highly effective hedging pairs for all but insignificant movements in the benchmark LIBOR rate, and that was consistently our experience for the years ended December 31, 2006, 2005 and 2004.
4. Regarding your fair value hedges of portfolios of municipal securities, please tell us the following:
· the characteristics you consider in aggregating securities that are deemed to be similar;
· the tenors of the interest rate swaps;
· how you consider the off-market components of the swap at the time of monthly de-designation and re-designation; and
· provide us with a specific example of how you assess effectiveness at inception of each monthly designated hedging relationship, include in your example how you consider all reasonably possible changes in fair value of the derivative and hedged items.
As we discussed in our response dated July 20, 2007 to Staff comment numbers 1 and 4 in the Staff’s letter of June 21, 2007, our use of derivative financial instruments substantially relates to our trading activities, and our use of derivative financial instruments in qualifying hedge accounting strategies is very limited.
We aggregated the municipal securities into portfolios based upon similar effective duration. Effective duration is defined as the weighted-average price change for a parallel shift in LIBOR yields in either direction. In general, duration is a weighted-average term-to-maturity of a security's cash flows, where the weights are the present values of each cash flow as a percent of the present value of all cash flows. A percentage change in price of a security is a direct function of its duration for small changes in yield. For example, a security with a modified duration of 9.00 would change approximately -0.90% for an increase in market yields of 10 basis points. An interest-rate swap agreement with a modified duration of 9.00 would change in an equal and opposite direction (+0.90%) for the same increase in market yields of 10 basis points. This relationship holds very closely for small, parallel changes in yields (as would be typical for a one-month-long horizon). Duration does not take into account non-benchmark factors such as credit risk. Importantly, the concept of matching effective durations is not dependent on “matching critical terms,” nor does it require that security values approximate par or that derivative fair values approximate zero. The concept holds at any price for a security or for a derivative as long as the duration is continuously recalculated based on the bond's (and the derivative’s) current price. Accordingly, matching effective durations works well for a dynamic strategy such as ours, and is not as vulnerable to the “law of small numbers” phenomenon as other more static strategies that may rely on matching of notional amounts and critical terms without periodic rebalancing.
We acknowledge that one deficiency in an effective duration matching strategy is that it presumes parallel LIBOR yield curve shifts. It is possible that yields on the LIBOR-based interest-rate swap agreements may not shift up or down by a similar amount in the event of a non-parallel yield curve shift, which could cause dissimilar changes in value of the interest-rate swap agreements and the municipal securities. However, since non-parallel yield curve behavior cannot be predicted, we believe that the use of effective duration captures all reasonable expected changes in yields, since it measures the average changes in price resulting from a shift in LIBOR yields in either direction.
For the periods presented, the tenors (hedge designation date through maturity date) for the interest-rate swap agreements ranged approximately from one month to seventeen years. The weighted-average tenor over this period, based on notional amount, was approximately four years.
3
This effective duration calculation, which we performed on both our hedging instruments and our hedged items, enabled us to assign instruments to portfolios based upon comparable interest-rate (LIBOR) sensitivity, rather than performing multiple interest-rate shock tests to establish an expectation of a highly effective hedge relationship. This sensitivity was based upon the effective duration by which we expected each individual item in the portfolio to respond in a generally proportionate manner to the overall change in fair value of the aggregate portfolio attributable to LIBOR (the hedged risk). The effective duration method demonstrated an alignment of the interest rate sensitivity of the derivatives and the securities and this was the basis for pairing derivatives and securities within these portfolios. The portfolio strata were pre-established and were not changed across hedge periods, although the individual interest-rate swap agreements and municipal securities allocated to each portfolio could have changed each month at each re-designation.
At each monthly re-designation, we acknowledged that the fair value of an interest-rate swap agreement may have been other than zero. However, the corresponding municipal security being hedged also would have had a fair value other than par. The effective duration hedging allowed us to determine the appropriate hedging relationship by aligning instruments based on market value. Because our hedge periods were short (only one month) and our effective duration calculations were updated each month as we contemplated our re-designations, our hedge effectiveness assessment data did not become stale with the passage of time or with the change in fair values of either our interest-rate swap agreements or our municipal securities.
The following example represents our effective duration calculation performed at the beginning of each hedge designation period and the impact of the results on the portfolio allocation, as well as the calculation of ineffectiveness at the end of the hedge period:
Bond Effective Duration Example:
Par: $5,000,000
Coupon: 5.00%
Maturity: July 1, 2015
Beginning yield: 4.070%
Beginning fair value: $106.30
Our hedge ratio at this point on the curve was 72.20%. The process is to shift the beginning yield by 1 basis point times the hedge ratio, or .007220% in this case. The process is to shift the curve up by this amount and value the bond, and then shift the curve down by this amount and value the bond.
Yield (curve shifted upward): 4.0772%
Price (curve shifted upward): $106.25
Yield (curve shifted downward): 3.9928%
Price (curve shifted downward): $106.35
4
Formula for Effective Duration:
(Price (curve shifted downward) - price (curve shifted upward)) / (2 * initial price * curve shift amount)
Our effective duration is therefore the following:
($106.35- $106.25) / (2 * $106.30* .00007220) = 6.51 years
Swap Effective Duration Example:
Notional: $3,800,000
Maturity: July 1, 2015
Coupon: 4.04%
Reset rate: 5.322%
For the swaps, we utilize a 1 basis point shift in yields for the effective duration calculation.
Beginning value: $381,708
Beginning price: $89.96
Fair value (curve shifted upward): $383,974
Price (curve shifted upward): $80.90
Fair value (curve shifted downward): $379,440
Price (curve shifted downward): $90.01
($90.01 - $89.90) / (2 * $89.96 * .0001) = 6.11 years
The specific effective duration portfolio for the instruments in the above example was 5.2 years to 7.4 years. In addition, the actual change in fair value of the derivative was $(23,139) and the corresponding actual change in fair value of the municipal security attributable to the hedged risk was $25,175, resulting in actual dollar-offset method ineffectiveness of $2,036. We noted that other instruments within this portfolio produced consistent results. This method was utilized historically across all reasonably possible changes in fair value of interest-rate swap agreements and municipal securities for monthly periods.
5
***
Please call the undersigned at 617-664-1110 or James J. Malerba, Senior Vice President and Corporate Controller at 617-664-8697, if you have any questions regarding the matters addressed in this letter or require any additional information.
| Sincerely, | |
|
|
|
|
| /s/ EDWARD J. RESCH |
|
| Edward J. Resch |
|
| Executive Vice President, Chief Financial Officer and Treasurer |
6
