The Affordability of Farmers

1
The Affordability of Farmers Markets in
Low-Income Boston Neighborhoods
Getting to the Root of Popular Perceptions
www.thefoodproject.org

• Robyn Lightner

The Food Project http//www.mainetoday.com/blogs/
inasnap/cat_farmers_market.html
2
Introduction

• By 2010, 21 farmers markets throughout the
  Boston metropolitan area participated in the BBB
  program and the usage of SNAP (41,402.28) along
  with BBB matches (36,409.46) combined for a
  total of 77,811.74 in sales- a 387 increase
  over the previous summer.
• Representing 28.5 of total Boston sales, the ten
  FM located in Dorchester and Roxbury reported a
  total of 21,699.58 in combined SNAP/BBB sales
  (The Food Project, preliminary report, 2011).
• BUT, there remains a widespread perception that
  farmers market produce is more expensive than
  conventional supermarket produce. This
  uninformed perception is a barrier to the
  attendance of price-sensitive shoppers at
  farmers markets.

3
Relative Cost of Direct Retail

• Problems with previous research did not devise
  plan to handle multi-leveled inconsistency of
  produce, combined produce into market basket,
  short time frame ignored effects of seasonality
  and volume.
• Demonstrated need for a replicable methodology
  for comparing produce prices at conventional
  grocery outlets with those at direct retail
  outlets
• Must be neighborhood-specific, include strategies
  that account for hard to measure effects of
  seasonality and quality

Pie Chart Comparing Prices at Seattle Farmers
Markets with Prices at QFC (SM)
Pirog and McCann, 2009 Kaemingk, 2010.
4
Observational Study

• Objective
• Clarify perception that farmers market produce is
  more expensive than conventional market produce
  in order to accurately educate communities about
  the local affordability of healthy food inform
  the efforts of community groups, policy
  advocates, farmers, etc.
• Learn how the price of produce differs
• between farmers markets and supermarkets
• in Dorchester and Roxbury
• Focus on specifics of a well-defined
• geographic/demographic area
• Assess relative quality between locations

5
Methods Variable Selection

• Neighborhoods
• North Dorchester, South Dorchester, Roxbury
• Similar socio-economic statistics, outlet
  density, momentum
• Retail Outlets
• All 10 FM and 7 SM in 8 mile area
• FM implemented for variety of reasons
• All outlets accept SNAP/EBT, WIC all FM accept
  FMNP and BBB
• Time Periods
• Eight 14-day periods between July 5 and October
  24, 2010
• Vendor Research
• Organic certification, time between point of
  harvest and sale, pricing strategies

Dorchester
Roxbury
Farmers Market
Supermarket
6
Farmers Market Information
7
Methods Variable Selection

• Produce Types
• Carrots unpeeled, with green tops and w/o green
  tops
• Cucumbers unwrapped, not English
• Onions yellow, large, loose (not bagged)
• Tomatoes (field) about fist size, not hot house/
  greenhouse, not heirloom, no vine
• Zucchini green summer squash
• White Potatoes loose
• Scallions green onions (not red bulb)
• Lettuce green leaf, not romaine/Bibb/iceberg
  variety, not bagged/washed/trimmed
• Bell Peppers (green) loose
• Green Beans loose

Bowdoin Street Health Center Robyn Lightner
8
Methods Data Collection

• All observations were taken within the same
  14-day period
• Recorded price/lb. data for the smallest, most
  conventional (i.e. non-organic) unit sold
• If non-pound unit, recorded unit price and took
  RS of unit weight to be converted into price/lb.
• Ranked overall visual quality of the produce item
  using a scale of 1 to 3
• All data collectors went through training
  workshop in order to ensure consistency
• Many problem-specific strategies

9
Description of Data

• Only 45 of possible observations at FM were
  actually realized and included in the final data
  set, compared with 88 of possible observations
  at SM.
• Both location types have fairly equal
  representation in the final data set
• An unrealized data point most commonly resulted
  from the case that a certain produce item was not
  available at the time or day of data collection,
  but may have also been the consequence of
  misreported data, a sold out item, or market
  cancellation due to weather, among others.
• The variance of mean price per pound at FM
  (0.904) was significantly lower (p .026) than
  the variance for the same value at SM (0.985),
  meaning that overall variation in mean price per
  pound was greater at SM.
• Mean quality at FM (2.74) was significantly
  higher (p lt 0.00) than at SM (2.46) by .27, or
  about 11.
• The standard deviation of quality rating was also
  lower at FM (0.47) than at SM (0.62), indicating
  that SM reported greater variation in quality.

10
Data Analysis
Multivariate Regression Model
Y lnPRICE  ?0 ?1 LOC ?2 TIME1 ?8
TIME7 ?9 PROD1 ?17 PROD9 ?18 QUAL1
?19 QUAL2 E

• How does the mean price per pound (PRICE) change
  with variation in location type (LOCATION), while
  produce type (PRODUCE), time period (TIME), and
  quality rating (QUALITY) are held constant?
• Results verified using log transformation to
  enhance normality of distribution
• Statistical analyses were performed using STATA,
  version 9.0 (StataCorp, College Station, TX,
  2008).

11
Regression Model
Regression Summary of lnPRICE on Successive Fixed
Effects

• A regression of the dependent variable lnPRICE on
  LOCATION controlling for PRODUCE, TIME, and
  QUALITY yields an estimated coefficient of .029
  and a p-value of 0.043. These results suggest
  that when produce type, time period, and quality
  are held constant, the price per pound is 2.9
  greater at farmers markets. The very low
  p-value of .043 asserts the statistical
  significance of this result.

12
Regression by Produce Type
13
Regression by Time Period
14
Regression by Time Produce Type
15
Regression by Time Produce Type
16
Other Discount Opportunities

• Bulk discounts
• i.e. Shaws loose onions were sold for 1.79/lb.
  compared with a 2 lb. package of onions for
  0.92/lb (about a 50 discount), and Tropical
  Foods sold loose potatoes for 0.79/lb. compared
  with a 5 lb. bag for 0.34/lb (nearly a 60
  discount).
• Value card discounts
• i.e. Stop and Shop Card offered zucchini for
  0.50/lb. (regular price 1.49) or green beans
  for 0.20/lb. (regular price 1.69).
• Bounty Bucks
• Up to 10.00 in free produce for SNAP users

17
More on Quality

• Organic or IPM Growing Methods
• 2 vendors certified (Langwater and Serving
  Ourselves), all others practice mostly organic
  growing methods
• Freshness
• All produce was grown within 2 hours driving time
  of Boston
• Likely sold within 24-48 hours of harvest

Examples of Comments on Quality
18
How Much Really is 3?

• The Economic Research Service of the USDA
  recently announced that an adult on a
  2,000-calorie diet could satisfy recommendations
  for fruit and vegetable consumption in the 2010
  Dietary Guidelines for Americans at an average of
  2 to 2.50 per day (Stewart et al., 2008).
• Using a mean of 2.25, an increase by 3 would
  increase daily cost of consuming the recommended
  F V amount by about 7-cents to 2.32.
• Over the course of the 16-week/112 day summer
  market season used in this study, purchasing the
  recommended daily amount of produce would
  increase from 252.00 to 259.84- just under
  8.00 over 4 months.
• This added expense could be more than compensated
  for in just one day using Bounty Bucks!

19
How can this information be used to educate
consumers about the affordability of neighborhood
farmers markets?

• Focus on the seasonal height of specific produce
  types and use time-sensitive publicity to
  encourage purchasing when price is lowest. For
  example, highlight bell peppers and cucumbers as
  Top Pickings or Highest Seeds during the last
  two weeks of September (time period 6), and do
  the same for scallions and lettuce during the
  first two weeks in October (time period 7).
• Create a citywide graphic that shows a visual of
  how much produce 20 could get a SNAP user at
  farmers markets and how that produce translates
  into a few days worth of vegetable servings.
• Create a side-by-side visual comparison of the
  amount of produce a shopper could buy using 10
  worth of SNAP at a supermarket, then show that
  same amount doubled for farmers markets to
  demonstrate the double voucher benefits available
  through Boston Bounty Bucks.

20
Special Thanks

• Cammy Watts, The Food Project
• Cathy Wirth, Bowdoin Street Community Health
  Center
• Professors Kate Sims, Jan Dizard, and Amy
  Wagaman, Amherst College
• Phuong Luong and Cara Brumfeld, The Food Project
• Nebi Stephens, Cynthia Loesch, Tim Deihl, Dave
  Dumaresque, The B.O.L.D. Teens of Codman Square

http//www.boldteens.org/media.html
21
Questions?