| Course | GF582 Statistical Methods for Decision Making |
|---|---|
| Unit | Unit 8 |
| Paper type | Regression output memo |
| Length | About 780 words, 5 pages double-spaced |
| Format | APA 7 student paper |
| School | Purdue University Global |
| Program | MS in Finance |
| Updated | October 2026 |
Free sample paper for GF582 Unit 8
$1.28 Million a Month per Tenth of a Point: A Unit 8 Regression Output Memo on Deposit Pricing
[Student Name]
Purdue University Global
GF582: Statistical Methods for Decision Making
Unit 8 Regression Output Memo
[Instructor Name]
[Date]
The bank, its competitors and every figure are composite and illustrative; the articles and textbook cited are real.
Memorandum
To: Treasurer From: Financial Analyst, Asset-Liability Management Date: May 5, 2026 Re: What the deposit data say about pricing the money market account
Recommendation
Do not raise the posted rate on the money market account. Instead, run a second three-month promotional rate on new money only, which the data suggest brings in deposits at a fraction of the cost.
The Model
Monthly net new balances, in millions of dollars, were regressed on two variables over the last 48 months: the gap between the bank's posted rate and the top local competitor's, in percentage points, and a flag for the eight months in which the bank ran a new-money promotion (Wooldridge, 2020).
Gap coefficient: 12.8, standard error 3.1, t-statistic about 4.13. Each percentage point of rate advantage over the top competitor is associated with about $12.8 million more net new money a month, or about $1.28 million for each tenth of a point.
Promotion coefficient: 4.6, standard error 1.4, t-statistic about 3.29. Promotional months brought about $4.6 million more net new money than other months with the same gap.
Intercept: 0.85. With the bank's rate equal to the competitor's and no promotion, balances grew about $0.85 million a month.
Fit: R-squared about 0.47; residual standard error about $2.9 million a month.
Converting the Coefficients Into a Decision
The bank's posted rate is now 4.05 percent and the top competitor's 4.30 percent, a gap of negative 0.25 points. The model predicts net outflows of about $2.35 million a month at that gap.
Raising the posted rate by 0.10 points would, by the model, add about $1.28 million a month. But the higher rate would apply to all $420 million of existing balances, costing about $420,000 a year in extra interest. Over a year, the new money would accumulate to about $15 million, averaging about $8 million. If the bank earns about 0.45 points more on deposits than on its next cheapest funding, wholesale advances, the new money is worth roughly $36,000 a year, against $420,000 of cost.
A three-month promotion on new money only would, by the model, add about $4.6 million a month above normal, about $13.8 million in total. The higher rate applies only to the new money, costing well under $100,000 a year even if the promotional rate is half a point above the posted rate.
The gap coefficient is large and clearly significant, but raising the posted rate pays the higher price on every existing dollar to attract a few new ones, so the coefficient that matters for the decision is the promotion flag, which buys new money without repricing the old.
Cautions
Correlation, not proof of cause. Competitors set their rates in response to the same market conditions, so the gap is not randomly assigned. Months when the bank's gap was favorable may have coincided with months when competitors were constrained for other reasons.
Omitted variables. The model does not include changes in the Federal Reserve's policy rate, seasonal patterns in business deposits or local economic news, any of which could affect balances. Adding a policy-rate variable left the gap coefficient similar, at 11.9, which is reassuring.
Serial correlation. Monthly deposit flows tend to be correlated over time, which can make standard errors too small. A check of the residuals showed modest correlation; correcting for it raised the gap coefficient's standard error to about 3.6, still leaving it significant.
Why Depositors Respond Slowly
Research on bank deposits finds that many depositors move money only when rate gaps become large, which gives banks market power over existing balances and explains why posted rates trail policy rates (Drechsler et al., 2017). That is consistent with this model: existing depositors stay even when the bank trails competitors, while new money responds to targeted offers.
What the Treasurer Can Tell the Board
The board asked last quarter why the bank's posted rate trails the leading competitor. The answer the data support is that matching the competitor would cost about $1 million a year in extra interest on existing balances, while bringing in deposits worth far less than that. Targeted promotions offer a cheaper way to grow balances when the bank needs funding.
Next Steps
Approve a three-month new-money promotion at 4.55 percent beginning next month, and rerun the model after it ends to test whether the promotion effect holds (Anderson et al., 2020).
References
Anderson, D. R., Sweeney, D. J., Williams, T. A., Camm, J. D., & Cochran, J. J. (2020). Statistics for business and economics (14th ed.). Cengage.
Drechsler, I., Savov, A., & Schnabl, P. (2017). The deposits channel of monetary policy. The Quarterly Journal of Economics, 132(4), 1819-1876. https://doi.org/10.1093/qje/qjx019
Wooldridge, J. M. (2020). Introductory econometrics: A modern approach (7th ed.). Cengage.
The Purdue Global GF582 Unit 8 instructions
Unit 8 of GF582 asks for a memo interpreting regression output for a decision maker. Lead with the recommendation. Describe the model briefly: the outcome, the predictors, the period and the number of observations. Interpret each coefficient in the units of both variables and scale it to the decision, such as dollars per tenth of a point. Convert the coefficients into the costs and benefits of the options available, including effects the regression does not capture directly, such as repricing existing balances. Report fit and uncertainty plainly. Explain cautions about causation, omitted variables and serial correlation, and any checks you ran. Connect the findings to research on the behavior involved. End with next steps that test the decision. Cite an econometrics text and relevant research in APA 7.
How the GF 582 Unit 8 example is put together
The memo leads with a recommendation against raising the posted rate. The model section reports the gap coefficient of 12.8, the promotion coefficient of 4.6, the intercept and fit, citing Wooldridge. A conversion section computes predicted outflows at today's gap, the cost of a 0.10-point increase on $420 million of balances and the value of the new money, then the cost of a three-month promotion. The highlighted sentence explains why the promotion flag, not the gap, drives the decision. Cautions cover non-random rate gaps, omitted policy-rate changes and serial correlation, each with a check. A section on depositor behavior cites Drechsler, Savov and Schnabl, a note gives the board an answer, and next steps cite Anderson and colleagues.
Where the points sit in the GF582 Unit 8 rubric
GF582 regression memos earn most of their points by translating output into decisions correctly. Graders check that each coefficient is interpreted in the right units and scaled to the choice at hand. The largest analytical share rewards converting coefficients into costs and benefits, including consequences outside the regression such as paying a higher rate on existing deposits. Plain reporting of fit and uncertainty is expected. Instructors in the MS in Finance program value cautions about causation and omitted variables, especially when checks are reported. A recommendation that follows from the economics rather than from significance alone, plus memo format and APA 7 references, completes the rubric.
GF582 Unit 8 help from the desk
Many GF582 Unit 8 memos list coefficients and p-values without saying what they mean in dollars. Others recommend the option tied to the largest or most significant coefficient without costing it. Effects outside the regression, such as repricing every existing balance, are ignored. R-squared is presented as a measure of whether the decision is right. Causation is assumed from observational data. The memo buries its recommendation. For any regression your case provides, follow the same order: recommendation, model, coefficients in decision units, costs and benefits, fit, cautions with checks and a next step that tests the result.
Get a GF582 Unit 8 example written to your instructions
The dataset or output from your Unit 8 assignment, the question it serves and the rubric are what the memo needs. A custom memo is written to the reader your prompt names, turns each coefficient into the units of that decision and states the model's limits, delivered within 24-48h and free for a first request. The paper above is an original model document written by our desk, not a submitted student paper and not an official Purdue University Global document.
More GF582 papers and MS in Finance samples
- GF582 Unit 1: Misleading Default Rate Discussion Post
- GF582 Unit 2: Bond Fund Descriptive Statistics
- GF582 Unit 3: Lending Probability Problem Set
- GF582 Unit 4: Card Portfolio Sampling Exercise
- GF582 Unit 5: Confidence Interval Workup in Two Parts
- GF582 Unit 6: Fund Persistence Seminar Reflection
- GF582 Unit 7: Scorecard Default Rate Hypothesis Test
- GF582 Unit 9: Model Comparison Exercise
- GF582 Unit 10: Statistical Decision Brief
- GF581 Unit 8: Cross-Border Capital Budgeting Analysis
- GF570 Unit 8: Foundation Attribution Analysis
- GF561 Unit 8: Asphalt Cross-Hedge Design Memo
- GF550 Unit 8: Retirement Withdrawal Order Memo
GF582 Unit 8 questions, answered
How should a coefficient be interpreted in the memo?
As the predicted change in the outcome for a one-unit change in that predictor, holding the others constant, in the units of both variables. Then scale it to a change the reader cares about, such as a quarter point rather than a full point, and attach the interval. A coefficient stated without units or scale rarely earns the interpretation marks.
Does a significant coefficient mean the predictor causes the outcome?
Not on its own. With observational data, a coefficient measures association after adjusting for the other predictors in the model, and anything left out can still drive both. Say which omitted factors are plausible, whether the predictor was set independently of the outcome, and what design, such as a controlled test, would support a causal claim.
Where does R squared belong in the memo?
Near the end, with a plain reading. It tells the reader how much of the outcome's month-to-month variation the model accounts for, not whether the decision is right. A modest R squared can still carry a precise, useful coefficient, and a high one can come from a model that answers the wrong question entirely.
Where can I find a free GF582 Unit 8 sample paper?
You can read a complete GF582 Unit 8 regression output memo on deposit pricing here, with coefficients converted to dollars, a cost comparison, cautions and checks, and APA 7 references.
Why not raise the posted rate if the gap coefficient is significant?
A higher posted rate applies to all $420 million of existing balances, costing about $420,000 a year, while the new money it attracts is worth only tens of thousands of dollars in funding savings.