Canonical page: https://sequoia-research.com/writing/delaware-statutory-trust-returns-why-similar-distribution-rates-can-produce-different-outcomes
Author: Sequoia Research
Published: 2026-10-17T10:33:31.899Z
Updated: 2026-10-17T10:33:31.899Z

# Delaware Statutory Trust Returns: Why Similar Distribution Rates Can Produce Different Outcomes

## What Is This Article About?

Two DSTs can send the same cash payment while giving their owners very different results over time. This article follows two invented investments that each pay 6% of the investor's original cash, then sell at different prices and dates. It shows how property costs, debt, reserves and the final sale affect the money an owner receives. It also tests a weaker sale and a household spending plan, so the comparison goes beyond the advertised rate.

## Table of Contents

- [What an Equal DST Distribution Rate Actually Tells You](#what-an-equal-dst-distribution-rate-actually-tells-you)
- [The Two Investments Behind the Same 6% Payment](#the-two-investments-behind-the-same-6-payment)
- [Follow Property Cash Through the Loan and Fees](#follow-property-cash-through-the-loan-and-fees)
- [How Sale Proceeds Change the Total Return](#how-sale-proceeds-change-the-total-return)
- [Compare Five Years With Seven Years](#compare-five-years-with-seven-years)
- [A Steady Distribution Can Hide a Reserve Draw](#a-steady-distribution-can-hide-a-reserve-draw)
- [Test a Lower Sale Price Without Changing the Payments](#test-a-lower-sale-price-without-changing-the-payments)
- [Separate Household Income From Remaining Capital](#separate-household-income-from-remaining-capital)
- [Check Whether the Two Advertised Rates Use the Same Basis](#check-whether-the-two-advertised-rates-use-the-same-basis)
- [Build a Comparison You Can Update](#build-a-comparison-you-can-update)
- [FAQ](#faq)

## What an Equal DST Distribution Rate Actually Tells You

An equal rate tells you the same amount of cash only when the two rates use the same dollar base and period. It says much less about the money that will remain when the investment ends.

Suppose an owner pays $200,000 for a DST interest. A $12,000 annual cash distribution is 6% of that payment. If a second DST also pays $12,000 on a $200,000 investment over the same year, the cash amounts match. That narrow statement is sound.

Several larger claims do not follow from it. The properties need not have the same value. The loans need not leave the same amount for owners at sale. One payment might come entirely from current operations while another uses cash set aside at closing. A sale can return more or less than the owner first paid.

It helps to keep three measures beside each other. The **distribution rate** divides a defined cash payment by a defined investment base. **Total cash received** adds all payments and final sale cash within the chosen period. **Net cash gain or loss** subtracts the owner's cash paid from that total. None of those measures, on its own, answers every investment question.

For example, receiving $12,000 a year for five years produces $60,000 of cash. If the final sale sends another $180,000, total receipts are $240,000. The nominal gain is $40,000 on the original $200,000. Calling the $60,000 of payments the complete profit would miss the $20,000 shortfall in the final check.

This is a cash comparison before personal taxes and inflation. It does not label any payment for tax purposes. A cash payment's economic source and its tax treatment require different records. A bank deposit alone does not settle either question.

The SEC advises investors to examine the method and included costs behind performance claims. It also distinguishes projections from actual results. That makes the status of each number as important as its size. [SEC: Performance Claims](https://www.investor.gov/introduction-investing/general-resources/news-alerts/alerts-bulletins/investor-bulletins-47)

All investments, rates, costs and outcomes below are invented teaching examples. They are not current offerings, historical sponsor results or forecasts. The point is to follow the same dollars from entry through exit, then change one part of that path at a time.

## The Two Investments Behind the Same 6% Payment

Consider Oak Terrace DST and Reed Crossing DST. In each example, a single class raises $20 million of investor equity. An investor named Luis contributes $200,000 and owns 1% of that class's economic interest. Every class receipt is shared in that proportion. There are no special payout tiers, other equity owners or later contributions in these models.

Both investments advertise a 6% annual distribution on original investor equity. For the main comparison, assume they actually pay that amount at each year-end until sale. This assumption lets us isolate the effect of the investment's other terms. It is not a claim that a target will be achieved.

Their entry budgets differ:

| Sources and uses at closing | Oak Terrace | Reed Crossing |
|---|---:|---:|
| Investor equity | $20,000,000 | $20,000,000 |
| Loan proceeds | $12,000,000 | $20,000,000 |
| Total sources | $32,000,000 | $40,000,000 |
| Property purchase price | $30,000,000 | $36,000,000 |
| Entry costs paid at closing | $1,000,000 | $3,000,000 |
| Opening cash reserve | $1,000,000 | $1,000,000 |
| Total uses | $32,000,000 | $40,000,000 |

Every dollar has one destination. The reserve remains cash in the trust. Entry costs have been paid and are no longer available to pay owners or buy property. The loan helps fund the purchase but must be dealt with at exit.

Luis's 1% share of Oak's purchase price is $300,000. Subtract his analytical share of the property loan, $120,000, then add $10,000 of entry costs and $10,000 of reserve cash. The result is his $200,000 contribution.

For Reed, the same bridge is $360,000 of purchase price, less $200,000 of debt, plus $30,000 of entry costs and $10,000 of reserve cash. It also totals $200,000. These loan shares describe exposure in the model. They do not establish Luis's personal liability for either loan.

Thus, equal investor checks do not buy equal gross property amounts or equal net property equity. Reed buys a higher-priced property with more debt and more entry costs. Oak starts with more net property equity per dollar Luis pays. That observation describes the budget; it does not prove which property will perform better.

Nor should the costs be subtracted from Luis's $200,000 again when calculating his return. They are already inside the funded budget. The SEC's fee guidance encourages readers to check both transaction and ongoing costs. The arithmetic still needs to count each charge once. [SEC: How Fees and Expenses Affect Your Investment Portfolio](https://www.investor.gov/introduction-investing/general-resources/news-alerts/alerts-bulletins/investor-bulletins/updated)

## Follow Property Cash Through the Loan and Fees

The property with the higher operating cash flow can still send the same amount to owners. Its loan and other cash costs may consume the difference.

In these models, net operating income, or NOI, means property income after the property costs included in the model, before loan payments and the separate trust expense shown below. All required property upkeep is included in those costs. There is no omitted work budget or separate annual reserve contribution. Actual offering records can use different definitions, so this definition belongs to these examples.

Assume the loans require interest only throughout the stated holds. Oak's $12 million loan costs $720,000 a year, or 6%. Reed's $20 million loan costs $1.3 million, or 6.5%. Principal stays unchanged. Each loan's assumed term covers its modeled sale date; neither example relies on a refinance or extension.

| Annual cash bridge | Oak Terrace | Reed Crossing |
|---|---:|---:|
| Property NOI | $2,000,000 | $2,600,000 |
| Loan interest | −$720,000 | −$1,300,000 |
| Separate trust expense | −$80,000 | −$100,000 |
| Cash paid to the investor class | $1,200,000 | $1,200,000 |
| Luis's 1% payment | $12,000 | $12,000 |
| Payment divided by his $200,000 | 6% | 6% |

Reed produces $600,000 more NOI. Its interest and trust expenses together also cost $600,000 more. The two differences cancel. Reading only the property income would miss the debt cost. Reading only Luis's payment would hide the different routes that produced it.

The numbers also explain why property price growth and investor results may diverge. Luis owns a claim on the cash remaining after the modeled obligations. The property can keep earning enough to fund his payments while the price available at sale leaves less capital for him.

We have assumed the annual cash is paid out in full. The original $1 million reserve earns nothing, receives no new cash and remains untouched in the main examples. It is added at sale once. This makes the reserve easy to trace and avoids counting it in both annual payments and final proceeds.

Before applying this bridge to a real DST, check the level at which a quoted cash figure is measured. A property's NOI, cash paid under a master lease, trust cash after expenses and an investor's bank receipt can be different amounts. If the reviewed documents contain an extra entity or contract, add that step. Do not force a more complex structure into this simplified path.

## How Sale Proceeds Change the Total Return

Oak sells after five full years for $33 million. Reed sells after seven full years for $36 million. The sale prices and dates are chosen assumptions, with no claim about the likely market value of either property.

Each sale incurs a total cost of 3% of its price. That charge represents all modeled sale expenses. The loan is repaid in full, and the unused $1 million reserve is released to the class. There are no other claims, unpaid fees, holdbacks or continuing assets. Final cash is paid immediately after the year-end operating payment.

| Final sale cash bridge | Oak, year five | Reed, year seven |
|---|---:|---:|
| Property sale price | $33,000,000 | $36,000,000 |
| Sale costs at 3% | −$990,000 | −$1,080,000 |
| Loan principal repaid | −$12,000,000 | −$20,000,000 |
| Unused reserve released | +$1,000,000 | +$1,000,000 |
| Final cash for investor class | $21,010,000 | $15,920,000 |
| Luis's 1% final cash | $210,100 | $159,200 |

Reed sells for the same price it paid for the property. Yet Luis's final check is $40,800 below his original contribution. Selling the property at its purchase price does not recover entry costs and sale costs. The loan and remaining reserve also belong in that calculation.

Oak's property rises by $3 million from purchase to sale. Luis's final cash exceeds his original contribution by $10,100. Its entry costs and sale costs use part of the property gain. The $10,000 reserve allocation returns cash he funded at the start; it is not a new gain created by the sale.

Now add the annual payments. Oak pays Luis $60,000 over five years and $210,100 at exit, for $270,100 in total. Subtracting $200,000 gives a $70,100 nominal pretax gain, or 35.05% of his contribution over the full hold.

Reed pays $84,000 over seven years and $159,200 at exit, for $243,200 in total. Its nominal pretax gain is $43,200, or 21.6% over the full hold. Reed sends more cash during ownership, yet less cash overall. Both pay the same 6% while Luis owns them.

The phrase “principal returned” can confuse this comparison. The whole final check is a cash receipt. It need not equal the original investment, and it may contain several economic sources. Add the actual final amount to the earlier receipts; do not add a second $200,000 as though the original investment were also repaid elsewhere.

## Compare Five Years With Seven Years

A five-year gain and a seven-year gain do not cover the same amount of time. Show that difference before calling either result better.

One useful measure is internal rate of return, or IRR. It finds a rate that balances the money paid out and received at each point in the schedule. FINRA describes IRR in present-value terms and notes limits, including assumptions tied to reinvestment and unsold values. Here, all final values are stipulated cash proceeds within the model, not estimates of unsold holdings. [FINRA: Regulatory Notice 20-21, Internal Rate of Return](https://www.finra.org/rules-guidance/notices/20-21)

With Luis's initial $200,000 outflow, equal year-end payments and the stated exits, Oak's modeled annual IRR is about 6.88%. Reed's is about 3.37%. Oak's final-year entry includes its $12,000 payment plus $210,100 of sale cash. Reed's final-year entry includes $12,000 plus $159,200. No payment is added twice.

Those rates describe the chosen schedules. They do not establish an available reinvestment rate after an early sale. They also do not turn either path into a forecast. A real dated record may have monthly payments, gaps and later final settlements, requiring a dated calculation rather than these equal annual periods.

A second comparison can use a shared year-seven endpoint. Suppose Luis saves all receipts without earning interest. Oak has produced $270,100 by year five, and that amount remains $270,100 at year seven. Reed produces $243,200 by year seven. The difference is $26,900. This simple comparison makes the treatment of Oak's earlier cash explicit.

A different assumption changes that endpoint. Suppose only Oak's $210,100 final check grows at a chosen 4% a year for the next two years. Assume no withdrawals and no extra fees or taxes in that hypothetical growth rate. It reaches $227,244.16. Add the $60,000 of earlier payments held without growth, and total year-seven resources are $287,244.16.

That extra $17,144.16 comes from the assumed next use of cash. Oak did not produce it during its ownership period. A fair record should label it separately, especially if the next investment carries different risks or costs. The 4% is a scenario input, not an available product quote.

If no next-use assumption is supported, leave it open. Luis can still compare each completed investment's receipts and dates. FINRA's return guidance emphasizes including costs and considering the holding period; it does not make a single percentage a substitute for those records. [FINRA: Calculating Your Investment Returns](https://www.finra.org/investors/insights/investment-returns)

## A Steady Distribution Can Hide a Reserve Draw

An unchanged payment does not prove that current property cash stayed unchanged. To see why, change only one year of Reed's main example.

Assume Reed's NOI falls to $2.2 million for that year. Loan interest remains $1.3 million and trust expense remains $100,000. Current operations leave $800,000 for the investor class, short of the $1.2 million scheduled payment.

For this separate branch, assume the governing terms allow a $400,000 reserve draw for the distribution, and the responsible party approves it. The draw brings that year's class payment to $1.2 million. Luis still receives $12,000. His bank record looks the same as in the main case.

The source record has changed. Luis's economic share of current operating cash is $8,000, and his share of the reserve draw is $4,000. The $1 million reserve now holds $600,000. The fact that the cash was available does not mean it can be spent again next year.

Assume all other years return to the original cash path, no reserve is rebuilt, and the sale terms remain unchanged. Reed's final class cash falls from $15.92 million to $15.52 million. Luis's final receipt becomes $155,200. His seven years of $12,000 payments still total $84,000, so full receipts become $239,200 and the nominal gain becomes $39,200.

The $4,000 difference from the main outcome is easy to lose in a review focused on the unchanged rate. It is the same $4,000 that moved from reserve cash into the earlier payment. Counting the original $1 million reserve at sale would wrongly create money that had already left.

The lower full-hold result comes from the operating shortfall. Using the reserve changes when existing cash is paid; it does not repair that shortfall. This does not classify the distribution as a return of capital for tax reporting. Nor does it assert that every DST can use a reserve this way. The permission is an express assumption in this model; a real review needs the actual reserve terms and approvals.

If only payment notices are available, write that the payments were confirmed but their funding remains unresolved. A conclusion about recurring cash needs more than proof that money arrived. Request the same-period operating cash bridge, reserve changes and any other sources used to fund the payment.

## Test a Lower Sale Price Without Changing the Payments

Return to Reed's original example, with the full $1 million reserve still intact. Now keep all seven years of $12,000 payments but lower the sale price to $30 million. This is a separate sale stress, not an addition to the reserve-draw branch.

Three percent sale costs are $900,000. Subtracting those costs and the $20 million loan, then adding the $1 million reserve, leaves $10.1 million for the class. Luis receives $101,000 at exit. Add his $84,000 of payments and his total cash is $185,000.

He has lost $15,000, or 7.5% of his original $200,000, in nominal pretax terms. The 6% payments arrived in every modeled year. They were real cash, but they did not cover the decline in final capital.

The property sale price fell 16.67% from $36 million to $30 million. The class's final cash fell from $15.92 million to $10.1 million, a decline of about 36.56%. That larger percentage reflects the fixed loan claim and the other terms of this example. It is not a universal multiplier for leveraged real estate.

A useful follow-up is the sale price needed just to avoid a nominal cash loss. Luis has already received $84,000, so his last check needs to be $116,000 to bring total receipts to $200,000. At 1% ownership, that requires $11.6 million for the class.

Under the same 3% sale cost, $20 million debt and $1 million reserve, that class amount requires a sale price of about **$31.55 million**. This threshold assumes all earlier payments arrived and ignores the time value of money, taxes and inflation. It is a cash break-even point, not a fair-value estimate or an acceptable return target.

A different question produces a different threshold. To receive the full original $200,000 in the final check alone, Luis would need $20 million of class sale cash. That requires a property sale near **$40.21 million** under these assumptions. Earlier payments do not reduce that second requirement because it concerns only the final check.

Write which threshold you mean. “Get my money back” can refer to total receipts, final capital or buying power after years of waiting. Leaving the phrase undefined can make two readers think they agree when they are testing different outcomes.

## Separate Household Income From Remaining Capital

Luis may spend his distributions rather than save them. That does not erase the investment's receipts, but it changes his household cash plan.

Assume he needs $12,000 a year from this part of his portfolio for seven years. For simplicity, all spending occurs at year-end. In both main cases, the annual DST payment covers that need while the investment remains open. Reed pays through year seven. Oak sells after year five, so its regular DST payments stop then.

If Oak's $210,100 final cash earns nothing, Luis can withdraw $12,000 in year six and another $12,000 in year seven. He ends year seven with $186,100. Reed has funded the same seven annual spending amounts and leaves him with its $159,200 final check. The difference remains $26,900.

Do not add Oak's earlier $60,000 to the remaining account balance in this household view. Luis already spent it. Those payments belong in the investment's return record, but they are no longer cash he can spend again. The same rule applies to Reed's $84,000 of consumed payments.

Now use a separate 4% growth assumption on Oak's final cash, with each later $12,000 withdrawal taken after that year's growth. The balance becomes $206,504 after year six and $202,764.16 after year seven. This is a new household scenario. It does not change Oak's original 6.88% modeled IRR.

The timing of expenses can matter even when annual amounts match. If a bill arrives before a payment, a year-end surplus does not pay the earlier bill by itself. A real spending plan should use the expected receipt dates, liquid cash available outside the DST and a delay case. This article's annual schedule does not prove month-by-month coverage.

A planned sale date also is not a withdrawal right. The SEC warns that private placements can be highly illiquid and may require an indefinite hold. Luis therefore needs a plan for a sale that happens later than either assumed date. A five-year model is a scenario to test, not a promise of cash in year five. [SEC: Private Placements Under Regulation D](https://www.investor.gov/introduction-investing/general-resources/news-alerts/alerts-bulletins/investor-bulletins/private-placements-under-regulation-d-updated-investor-bulletin)

## Check Whether the Two Advertised Rates Use the Same Basis

The main examples deliberately use the same original-equity base. Real comparisons first need to establish that the quoted rates mean the same thing.

Start with the dollars beneath the percentage. In a separate example, a 6% rate applied to $180,000 produces $10,800. If the investor actually paid $200,000, that amount is 5.4% of the cash paid. Another 6% rate applied to the full $200,000 produces $12,000. The printed rates match; the checks do not.

The smaller base might have a defined purpose in a particular report. The review should preserve that definition, then show the payment on a common investor-cash base. Do not silently rename it or conclude that the difference proves misleading conduct. First establish the exact wording, costs and calculation method.

Next, distinguish a current pace from a completed year. A $1,000 monthly payment has a simple annualized pace of $12,000. If a new investor received only nine such payments during the measured year, actual cash was $9,000, or 4.5% of $200,000. That does not, by itself, prove a payment cut. The missing months might precede the investor's ownership or the first payment date.

Record the investment date, first covered period and actual payment dates. A full calendar-year rate and a partial ownership-year total can both be correct while answering different questions. A first distribution that covers several months also should not be treated as the normal monthly amount without checking the period.

Then ask which deductions are inside the number. Oak's main annual payment already reflects the listed loan interest and trust expense. Subtracting those again from Luis's $12,000 would understate cash. A separate account fee paid from another source would need its own treatment if the intended comparison is after that fee.

The SEC's performance guidance calls attention to included and excluded fees and the method behind a presentation. Apply that principle at the level of the actual claim. A documented common rate is the start of a comparison; different bases, dates or fee scopes need to be reconciled before the two rates are used side by side. [SEC: Performance Claims](https://www.investor.gov/introduction-investing/general-resources/news-alerts/alerts-bulletins/investor-bulletins-47)

## Build a Comparison You Can Update

A useful comparison preserves a short cash trail for each investment. It should be possible to replace an assumption with a new record without rebuilding the entire story from memory.

Begin with the same investor contribution and the exact interest it buys. Attach the funded budget, the distribution-rate definition and the relevant class terms. Confirm that fees paid at entry are already part of that budget or identify any charges paid outside it. Keep debt in the property-to-equity bridge without adding it to the investor's cash contribution.

Next, save the operating bridge for each measured period. Show current cash, loan payments, trust costs, reserve changes and owner payments. A stated reserve balance should have an as-of date. If an earlier balance preceded a draw, it cannot support a later exit calculation without an update.

Keep a separate exit sheet with price, costs, loan payoff, remaining cash and other claims. Each field should say whether it is a forecast, a dated estimate or an actual closing amount. If part of the final settlement is still withheld, the paid amount and the unresolved amount should occupy separate lines.

For Oak and Reed, the few facts that change the conclusion are easy to name. Oak's model depends on its five-year sale and $210,100 final investor cash. Reed's depends on seven years of payments and $159,200 at exit. A reserve draw, lower price or later payment changes a particular line. An unchanged distribution rate does not confirm the rest of those assumptions.

When evidence is missing, keep the comparison unfinished in that area. A known distribution history does not fill in an unknown loan payoff. A purchase price does not establish a current sale value. The cleanest presentation can still contain an unresolved field, provided the reader sees how it limits the conclusion.

Finally, record what the comparison is meant to answer. A household seeking near-term cash has a different question from an owner comparing full-hold gains. Both can use the same records. Keeping income, total receipts, remaining capital and dates visible prevents one convenient percentage from standing in for all four.

## FAQ

### If two DSTs both pay 6%, are their returns the same?

Their cash distributions match only if the dollar base and measured period also match. Total results can differ because of exit proceeds, costs, loan balances and holding periods. Oak and Reed each pay $12,000 a year on $200,000 in the examples, yet their full receipts are $270,100 and $243,200. Those are model results, not market expectations.

### Can a DST pay every scheduled distribution and still lose money?

Yes, the cash arithmetic allows that outcome. In the separate Reed sale stress, seven years of payments total $84,000, but the final check is only $101,000. Total receipts of $185,000 fall $15,000 below the original contribution. A distribution record does not guarantee recovery of the remaining capital.

### Why does selling a property at its purchase price fail to return all equity?

The investment may have paid entry costs, and the sale may have costs of its own. Debt and remaining cash also affect the final equity amount. Reed's property sells for its original $36 million in the main example, but the class receives $15.92 million after the stated costs, loan and reserve. Trace the whole budget before drawing a conclusion from price alone.

### Does an unchanged distribution mean operating cash is stable?

No. A payment might use another allowed cash source. The separate reserve example keeps Reed's payment unchanged while shifting $400,000 out of its reserve. That same cash is then unavailable at sale. The model does not establish tax treatment or permission to use reserves in an actual DST; both need their own records.

### Should I choose the shorter holding period?

A shorter modeled hold is not automatically better and may not happen as planned. It makes cash available earlier only if the assumed sale and settlement occur. Compare the amount, timing and possible next use of that cash, along with risks and your needs. Do not assign an unsupported reinvestment return merely to make the two end dates match.

### Are spent distributions still part of investment performance?

Yes. Cash received remains part of the investment's return record even if the owner spends it. It should not also appear as unspent household savings. The article keeps those two views separate: total investment receipts measure the cash produced, while a spending plan tracks how much remains available.

### What should I request before comparing distribution rates?

Request the rate's dollar base, covered dates and calculation method; the current cash and reserve bridge; the funded entry budget; and the assumptions behind final proceeds. Identify included fees and any separate account costs. Mark projections and incomplete records plainly. A comparison becomes more useful when the two rates can be traced to dollars on the same terms.

## About Sequoia Research

Sequoia Research, LLC provides independent research on DSTs and other real estate investment offerings. This educational library explains investment records, comparison methods and the limits of available evidence.

## Sources

- [FINRA: Calculating Your Investment Returns](https://www.finra.org/investors/insights/investment-returns), December 4, 2024. Cost and holding-period context.
- [SEC: Performance Claims](https://www.investor.gov/introduction-investing/general-resources/news-alerts/alerts-bulletins/investor-bulletins-47), September 15, 2022. Calculation scope, costs and hypothetical versus actual results.
- [SEC: How Fees and Expenses Affect Your Investment Portfolio](https://www.investor.gov/introduction-investing/general-resources/news-alerts/alerts-bulletins/investor-bulletins/updated), July 23, 2025. Transaction and ongoing costs and disclosure sources.
- [FINRA: Regulatory Notice 20-21](https://www.finra.org/rules-guidance/notices/20-21), July 1, 2020. Narrow IRR definition and limitations; no claim about current advertising permission.
- [SEC: Private Placements Under Regulation D](https://www.investor.gov/introduction-investing/general-resources/news-alerts/alerts-bulletins/investor-bulletins/private-placements-under-regulation-d-updated-investor-bulletin), August 17, 2022; updated September 21, 2026. Illiquidity and uncertain holding periods.

Sources accessed October 17, 2026. All investment examples and outcomes are original hypothetical models.


## Related Reading and Tools

- [Delaware Statutory Trust Returns: Dated Cash-Flow Return Calculator](https://sequoia-research.com/writing/delaware-statutory-trust-returns-dated-cash-flow-return-calculator)
- [Delaware Statutory Trust Returns: Exit Assumption Sensitivity Grid](https://sequoia-research.com/writing/delaware-statutory-trust-returns-exit-assumption-sensitivity-grid)
- [Delaware Statutory Trust Fees: A Guide to Upfront, Ongoing, and Exit Costs](https://sequoia-research.com/writing/delaware-statutory-trust-fees-a-guide-to-upfront-ongoing-and-exit-costs)