April 17, 2015

Division of Corporation Finance

Securities and Exchange Commission

100 F Street, NE, Mail Stop 3720

Washington, DC  20549

Re: AT&T Inc.

Form 10-K for Fiscal Year Ended December 31, 2014

Filed February 20, 2015

File No. 001-8610

Ladies and Gentlemen:

We are in receipt of the letter from the staff (the "staff") of the Securities and Exchange Commission (the "Commission"), dated April 3, 2015 to Mr. John J. Stephens of AT&T Inc. (the "Company" or "we").  For your convenience, we have included the staff's comments herein and included our responses accordingly.

We acknowledge that the adequacy and accuracy of the disclosure in the filing is our responsibility.  We acknowledge the staff's position that comments or changes to disclosure in response to the staff comments do not foreclose the Commission from taking any action with respect to our filings.  We also understand that the staff comments may not be asserted as a defense in any proceeding initiated by the Commission or any person under the federal securities laws of the United States.

The Company's responses to the staff's comments are set forth below.

Form 10-K for Fiscal Year Ended December 31, 2014

Notes to the Consolidated Financial Statements

Note 12. Pension and Postretirement Benefits, page 64

1. We note your disclosure that you changed the method you use to estimate the service and interest components of net periodic benefit costs for pension and other postretirement benefits. You disclose that you have elected to utilize a full yield curve approach in estimation of these components. Please compare and contrast for us in greater detail the previous method with the current method, and explain to us how the current method complies with ASC 715-30-35 and ASC 715-60-35.

Response:

Under both our previously applied method and newly adopted method of accounting, our benefit obligation is estimated using individual discount rates (i.e., spot rates) associated with the corresponding expected benefit payments for each particular year.  As permitted by ASC 715-30-35-44, these individual discount rates are determined through the use of a yield curve comprised of multiple rates of return on several high-quality, fixed income corporate bonds available at the measurement date.  These discount rates reflect the hypothetical rate at which the benefit obligation could effectively be settled or paid out to participants. Using these individual rates, the projected benefit cash outflows for each particular year are discounted for the number of years from the measurement date. Each of the discounted cash outflows is added together to determine the total projected benefit obligation.  The only change resulting from adoption of the full yield curve approach is the determination of the interest rate (i.e., the unit of accounting) used to accrue service and interest cost.

In connection with the required interim remeasurement of our pension obligation as of October 1, 2014, we adopted a change in the method of estimating the service and interest cost components of net periodic benefit cost.  We determined that this newly adopted method is consistent with the requirements of ASC 715 and results in a more precise determination of the effective settlement rates and calculation of the service and interest cost components of net periodic benefit cost because it:

·Derives those cost components from the specific annual spot rates along the spot yield curve rather than using a single weighted-average rate, which we historically used and is widely used in practice by other entities, and

·Eliminates actuarial gains and losses created solely as a result of the use of the previously applied method (i.e., use of individual spot rates to calculate the benefit obligation while using a single weighted-average rate for interest and service costs results in actuarial gains or losses).

Below is a comparison of our previously applied method of accounting and the full yield curve approach when determining the components of our net periodic benefit cost.

Previously applied method of accounting

Using that discounted benefit obligation noted above, we have historically determined a single weighted-average discount rate for measuring the associated service and interest cost components of net periodic benefit cost over the subsequent year.  Under our previously applied method, the assumed discount rate used to measure the service and interest cost components was the single weighted-average rate derived from the present value of the benefit obligation and the expected cash outflows.  We believe that our historical application of a single weighted-average discount rate was an appropriate and acceptable method determined in accordance with ASC 715-30-35-45 and one that has been and continues to be widely used in practice by other entities.

Newly adopted method of accounting

In an effort to more precisely measure the service and interest cost components, we made the decision to utilize a "full yield curve" approach.  Under this approach, and consistent with our measurement of the benefit obligation, service cost is computed by applying the specific spot rates along the yield curve that correspond to the timing of each service cost cash flow.  Because the service cost component relates to the active participants in the plan, the relevant cash flows on which to apply the yield curve are considerably longer in duration on average than the total projected benefit obligation cash flows which also include benefit payments to retirees.  Similarly, interest cost is computed by multiplying annual spot rates by the corresponding annual discounted cash flows.

Comparison of methods

The use of the full yield curve approach results in a more precise measurement of the service and interest cost components of net periodic benefit cost.  In comparison to our historical approach, the full yield curve approach eliminates any gains and losses based upon interest rate trends (e.g., built-in gains in interest cost in an upward sloping yield curve scenario), or gains and losses merely resulting from the timing and magnitude of the Company's cash outflows associated with its benefit obligations.  Such results can occur under our previously applied method that used a single weighted average discount rate as a substitute for a more precise series of discount rates, even if the estimated rates underlying the benefit obligation calculation were unchanged.

Because we follow a policy of immediate recognition of actuarial gains and losses on pension and other postretirement obligations in the period of remeasurement, the full yield curve approach will not affect the annual measurement of our total benefit obligations or annually reported net periodic benefit cost.  Rather, it only results in a more precise measurement of the various components of the interim and annual net periodic benefit cost: (i) service cost, (ii) interest cost and (iii) actuarial gain/loss (with these components being separately disclosed in the notes to the consolidated financial statements).  Under the current interest rate environment, this change is expected to reduce the service and interest cost components that are recorded throughout the year with an offsetting reduction in actuarial gains upon remeasurement.  In an upward sloping yield curve scenario, we expect interim reporting periods that do not require a remeasurement to have lower net periodic benefit cost as compared to the amounts computed under our previously applied method.

In regard to our evaluation of how the full yield curve approach complies with authoritative guidance (i.e., ASC 715-30-35 and 715-60-35), please refer to our response to staff comment #3 noted herein.

2. Please describe to us the impact the change had on your GAAP and non-GAAP results for the year and quarter ended December 31, 2014. In addition, describe the impact you expect the change to have on your quarterly and annual results for 2015 on a GAAP and non-GAAP basis.

Response:

There was no impact of the change on our consolidated GAAP results for the year and quarter ended December 31, 2014.  Our segment operating results were not significantly impacted.  On a non-GAAP basis, due to the resulting decreased service and interest costs and increased actuarial loss (which is adjusted out of certain of our non-GAAP measures), this change impacted our adjusted operating income and margin, our adjusted EBITDA, and our adjusted diluted earnings per share (EPS).  For the year and quarter ended December 31, 2014, this change resulted in a $150 million increase in adjusted operating income and adjusted EBITDA, an 11 basis points and 44 basis points increase in operating margin for the respective year and quarter, and a $0.02 increase in adjusted diluted EPS.  Because the interest cost and actuarial loss components of net periodic benefit cost are excluded from our segment operating results, the change had an insignificant impact on our non-GAAP segment results, increasing adjusted segment EBITDA by approximately $4 million for our Wireless segment and $16 million for our Wireline segment.

For the full year 2015, we expect the change to have no impact on our consolidated GAAP results, as the expected reduction in service and interest cost and corresponding increase in actuarial gains and losses will not change our annual net periodic benefit cost.  Our segment operating results for the year are not expected to be significantly impacted by the change.  Absent the need for any interim remeasurement of the Company's benefit obligations, we estimate the impact of the change on our consolidated GAAP results for each 2015 interim reporting period to be a reduction of net periodic benefit cost of approximately $260 million, resulting in an increase to net income of $160 million.  On a non-GAAP basis, the change is expected to increase adjusted operating income and adjusted EBITDA by approximately $1.0 billion, increase adjusted net income by approximately $650 million and increase adjusted diluted EPS by approximately $0.11 for the year ending December 31, 2015.

3. We note that you have accounted for the change as a change in accounting estimate effected by a change in accounting principle. Please explain to us your basis for accounting for the change in this manner, rather than a change in estimate or a change in accounting principle, as defined in ASC 250-10.

Response:

We concluded that adoption of the full yield curve approach was a change in accounting estimate that is effected by a change in accounting principle, as defined in ASC 250-10-45-18, and accordingly accounted for the preferable change prospectively.  The effect of the change in accounting principle, or the method of applying it, is inseparable from the effect of the change in accounting estimate.  The use of the full yield curve approach is a change in the method used to calculate the assumed discount rate to estimate the service and interest cost components of net periodic benefit cost, while utilizing the same information as under our previously applied method.  That is, there is no change in the underlying information used to estimate the service and interest cost.

As identified in ASC 250-10-45-18 "distinguishing between a change in an accounting principle and a change in an accounting estimate is sometimes difficult."  When evaluating this change, we concluded that it is similar to changing the depreciation method for a long-lived, nonfinancial asset, which the FASB cited as an example of a change in estimate effected by a change in accounting principle.  Specifically, ASC 250-10-45-19 explains that an entity may be justified in making a change in estimate effected by a change in accounting principle when it concludes that the pattern of consumption of the expected benefits of an asset has changed, and it determines that a new depreciation method better reflects that pattern.  Similarly, we concluded that the change to the full yield curve approach was made as part of the continuing process of using the best available information and methods to apply that information to reflect our estimate of the service and interest cost components of the Company's net periodic benefit costs.  We considered this change to be similar to a change from the use of a composite depreciation method to the use a component depreciation method, with both resulting in more precise computations. Specifically, our change moves from a computation that accretes the entire benefit obligation using a weighted average factor to individual computations for each annual payment obligation.

We also note that accounting for this change as a change in estimate effected by a change in accounting principle results in increased transparency to our investors and other users of our financial statements compared to accounting for the change solely as a change in estimate. As a result of concluding that this change is a change in estimate effected by a change in accounting principle, we were required to conclude that the change is preferable, include a preferability letter from our independent registered public accounting firm and include additional disclosures in our financial statements. These items would not have been communicated to our investors or other users of our financial statements had we determined that the change is a change in estimate. For a company that immediately recognizes actuarial gains or losses, concluding that the change is a change in estimate would not have resulted in any changes to the balance sheet, income statement and statement of cash flows when compared to a change in estimate effected by a change in accounting principle.

We also considered whether the change was a change in accounting principle. While the same underlying information is used for either our historical or our current estimate of the service and interest cost components of our benefit obligations, we do not believe this change reflects only a change in accounting principle. This change represents a change in the level (i.e., unit of accounting) at which the information is applied to complete a more refined estimate. Because we are on an immediate recognition approach, the change has no effect on our annual net periodic benefit cost and there would not be an effect to apply retroactively other than a reallocation between the service and interest cost components with an offset to actuarial gains or losses.  We believe this change is a refinement of our previous estimate (i.e., use of a more precise calculation). Similarly, a company may refine its estimate of depreciation expense and change from estimating depreciation using the composite method to the component method. As noted previously, such a change from the composite method to the component method is considered a change in estimate effected by a change in accounting principle (see ASC 250-10-45-18).

4. Please describe to us in detail your basis for determining the change is preferable. Refer to ASC 250-10-45-19.

Response:

In evaluating the preferability of the change in accordance with ASC 250-10-45, we primarily considered the three following factors: 1) authoritative support for the change, 2) the rationality of the change, and 3) whether the change is supported by current industry practice.  Our evaluation of each of these three factors, and our basis for concluding the change was preferable, is summarized as follows:

Authoritative Support

ASC 715-30-35-43 to 45 (other postretirement plans: ASC 715-60-35-79 to 82) provides guidance on the determination of applicable discount rates used in the measurement of the projected benefit obligation and accumulated postretirement benefit obligation as well as the estimation of the service cost and interest cost components of net periodic benefit cost:

35-43: Assumed discount rates shall reflect the rates at which the pension benefits could be effectively settled. It is appropriate in estimating those rates to look to available information about rates implicit in current prices of annuity contracts that could be used to effect settlement of the obligation (including information about available annuity rates published by the Pension Benefit Guaranty Corporation). In making those estimates, employers may also look to rates of return on high-quality fixed-income investments currently available and expected to be available during the period to maturity of the pension benefits. Assumed discount rates are used in measurements of the projected, accumulated, and vested benefit obligations and the service and interest cost components of net periodic pension cost.

35-44: The preceding paragraph permits an employer to look to rates of return on high-quality fixed-income investments in determining assumed discount rates. The objective of selecting assumed discount rates using that method is to measure the single amount that, if invested at the measurement date in a portfolio of high-quality debt instruments, would provide the necessary future cash flows to pay the pension benefits when due. Notionally, that single amount, the projected benefit obligation, would equal the current market value of a portfolio of high-quality zero coupon bonds whose maturity dates and amounts would be the same as the timing and amount of the expected future benefit payments. Because cash inflows would equal cash outflows in timing and amount, there would be no reinvestment risk in the yields to maturity of the portfolio. However, in other than a zero coupon portfolio, such as a portfolio of long-term debt instruments that pay semiannual interest payments or whose maturities do not extend far enough into the future to meet expected benefit payments, the assumed discount rates (the yield to maturity) need to incorporate expected reinvestment rates available in the future. Those rates shall be extrapolated from the existing yield curve at the measurement date. The determination of the assumed discount rate is separate from the determination of the expected rate of return on plan assets whenever the actual portfolio differs from the hypothetical portfolio described in this paragraph. Assumed discount rates shall be reevaluated at each measurement date. If the general level of interest rates rises or declines, the assumed discount rates shall change in a similar manner.

35-45: Interest rates vary depending on the duration of investments; for example, U.S. Treasury bills, 7-year bonds, and 30-year bonds have different interest rates. Thus, the weighted-average discount rate (interest rate) inherent in the prices of annuities (or a dedicated bond portfolio) will vary depending on the length of time remaining until individual benefit payment dates. A plan covering only retired employees would be expected to have significantly different discount rates from one covering a work force of 30-year-olds. The disclosures required by SubTopic 715-20 regarding components of the pension benefit obligation will be more representationally faithful if individual discount rates applicable to various benefit deferral periods are selected (emphasis added).A properly weighted average rate can be used for aggregate computations such as the interest cost component of net pension cost for the period.

While ASC 715-30-35-45 notes that an entity can use a weighted average rate when computing its service and interest cost components, the full yield curve approach measures the service and interest cost components by multiplying annual spot interest rates (the same rates that would be utilized to calculated weighted average rates) by the corresponding annual discounted cash flows and summing the result.  This approach is consistent with the bolded sentence above in ASC 715-30-35-45 that advises that applying individual period rates will be more faithful to the required disclosures.  Specifically, the resulting calculations from the full yield curve approach are more precise as individual discount rates are applied to the various benefit deferral periods instead of a single equivalent or weighted-average rate.  Additionally, this approach will not impact the annually reported net periodic benefit cost given the Company's policy to recognize all actuarial gains and losses immediately into earnings upon remeasurement of the benefit obligations, which occurs as of December 31 of each year or upon an interim remeasurement event (e.g., plan changes, settlements or other events significantly affecting the recorded obligations).  As for interim financial statements, the net periodic benefit cost will be impacted by the change because only the service cost and interest cost components of net periodic benefit cost are recognized. However, as explained further herein, the full yield curve approach results in a more precise reflection of the interim net periodic benefit cost as it eliminates the actuarial gains or losses that may result under our previously applied method.

Rationality

Consistent with our historical accounting, we estimate the expected future cash outflows for each future year.  The Company utilizes a spot rate yield curve to discount each future year's expected cash outflow at the corresponding spot rate for the various annual intervals.  The benefit obligation for our pension and other postretirement benefit plans is determined by summing the discounted future cash outflows in a manner consistent with the guidance in ASC 715-30-35-29 for pension benefits and ASC 715-60-35-41 for other postretirement benefits.  Regarding the application of the full yield curve approach, the measurement of the service and interest cost components is determined as follows:

·Measurement of Service Cost - In order to measure the service cost, we now apply assumed discount rates from the spot yield curve to the service cost cash flows in the same way that we apply the spot yield curve to the projected benefit obligation cash flows to arrive at the projected benefit obligation.  The same process also applies to the accumulated benefit obligation cash flows.  This view is consistent with ASC 715-30-35-43 and ASC 715-30-35-29 which note the commonality of the measurements of the service cost, the accumulated benefit obligation, and the projected benefit obligation.  Since the service cost component relates to the active participants in the plan, the relevant cash flows are, on average, much longer in duration than the total projected benefit obligation cash flows which include the benefit payments to retired participants.  This longer duration has the effect of a higher level equivalent yield for the cash flows related to active participants as compared to the total participants in the plan, assuming an upward sloping yield curve.

·Measurement of Interest Cost - In order to measure the interest cost component for the current period, we estimate the interest cost using each assumed discount rate in the spot yield curve.  Specifically, this component is measured by multiplying annual spot rates by the corresponding annual discounted cash flows and summing the result.  Under our historical approach, the weighted-average discount rate was used to measure the interest cost.

When our historical policy is compared to the full yield curve approach described herein, the resulting differences principally exist because of the weightings attributed to the individual spot rates, which consequently affects the interim and annual measurement of the service and interest cost components and the recognized actuarial gain or loss upon remeasurement of the benefit obligations.  The magnitude of the change in interim periods will depend on the shape of the spot yield curve at the most recent measurement date.  While the resulting amounts for the individual components of net periodic cost would differ under the two approaches, the total net pension and other postretirement benefit cost recognized through earnings would be the same in an annual period because of the Company's policy to recognize all actuarial gains and losses immediately into earnings upon remeasurement of the benefit obligations.  As a result of the change in method of calculating service and interest cost, those calculations are now more consistent with our calculation of the benefit obligation and result in more accurate interim reporting and measurement of the components of our net periodic benefit cost (i.e., the full yield curve approach eliminates the actuarial gains or losses that may result under our previously applied method).

Industry Practice

It is important to note that the full yield curve approach will not impact the Company's annually reported net periodic benefit costs — that is, the benefit obligations will continue to be measured using the spot yield curve approach and the total annual net periodic benefit cost will not change given the Company's policy of immediate recognition through earnings of actuarial gains and losses for our pension and other postretirement benefit obligations.  As a result, comparability would be preserved with other entities following immediate recognition through earnings of actuarial gains and losses with respect to total benefit obligations and the related net periodic benefit cost in annual periods (even though the calculation of the components may differ).  We believe that the full yield curve approach, while not prevalent in practice, is a more precise measure of the components of the Company's net periodic pension and other postretirement benefit cost.  This approach is also supported by our outside actuaries who have extensive experience in the field of benefit plan measurement.

Based on the supporting authoritative guidance and relevant facts and circumstances as discussed above, we believe the accounting change is preferable because it results in a more precise measurement of service and interest costs.Please let me know if you have additional questions or comments.

Sincerely,

/s/John J. Stephens

John J. Stephens

Senior Executive Vice President

And Chief Financial Officer