Finding Fraud

videoType

dynamic-media

videoDmUrl

keyFrameImage

timestamp

00:01 Great.

00:02 Good morning,

00:03 good afternoon and good evening,

00:06 everyone.

00:06 It's great to see so many of you connected.

00:09 We're at 78 attendees this morning,

00:12 uh,

00:12 afternoon.

00:13 Great that you were able to,

00:14 to join us.

00:15 This is the latest,

00:16 uh,

00:17 BBB Round Bag brunch or

00:19 breakfast or dinner wherever you are,

00:21 uh,

00:21 from the GovTech series,

00:23 and we're trying to vary our time so that we can have more attendance from the field.

00:27 Our topic today is the use of technology to

00:30 find fraud and corruption in the public administration.

00:33 The United Nations estimates that $1 trillion is paid in bribes,

00:38 $2.6 trillion lost to corruption

00:41 every year,

00:42 and combined this is 5% of global GDP.

00:46 In low income countries,

00:48 the cost of corruption

00:49 is estimated to be about 10 times

00:52 the inward flows from ODA.

00:55 In the current environment,

00:56 COVID-19 has caused a greater perception of corruption being a problem,

01:02 and we've seen lower

01:04 trust in government institutions.

01:06 So this new report is even more relevant today.

01:09 Finding fraud,

01:10 govtech,

01:11 and fraud detection in the public administration

01:13 examines new approaches to using technology,

01:16 digital analysis of data,

01:18 and artificial intelligence to detect and prevent

01:20 fraud and corruption in the public administration.

01:23 The focus of the report is on government systems for procurement,

01:27 financial management,

01:28 and human resource management.

01:30 The report also addresses other legal,

01:33 policy and political requirements for these tools to be successful,

01:37 and the report's intended to be a practical guide for practitioners,

01:40 policymakers and government officials,

01:42 and for our TTLs to use in their policy dialogue.

01:45 This morning we will have opening remarks from

01:48 the governance GP Global director Ed Ololo Okeri,

01:51 and then we'll hear from two of the experts on the report,

01:54 Hunt Lacasia,

01:55 a senior procurement specialist,

01:57 and Iza Malik,

01:58 a public sector specialist in the Menda region.

02:01 And then we'll turn to our discussants this morning.

02:03 So a warm welcome to Irina Smeliova and Donna Andrews,

02:07 our discussants,

02:08 and we will end with some Q&A from the panel.

02:11 Please use the Q&A chat box and post your question there,

02:16 and we will come to you at the end so you can post questions as we go through.

02:20 Ed,

02:20 over to you for welcoming remarks.

02:23 OK,

02:24 thanks,

02:24 uh,

02:25 Tracy.

02:25 Uh,

02:26 good morning,

02:26 good afternoon and good evening to everyone.

02:30 Uh,

02:30 I'm delighted to join you in this event that is focused on

02:35 the important issue of fraud.

02:37 This is an age-long issue that continues to be with us,

02:41 even in a period of crisis as we are currently.

02:46 Over the years,

02:48 various approaches and tools have been developed to prevent and detect fraud.

02:54 This mostly build

02:56 on rule of law

02:58 and transparency as well as the

03:00 classical internal and external control principles.

03:05 In recent years,

03:07 breakthroughs in digital technologies have expanded

03:11 the horizon of possibilities

03:13 and provided an array of new tools

03:17 to governments to help them tackle fraud challenges

03:21 and also address the broad issue of controlling corruption.

03:26 GovTech,

03:27 including e-government systems

03:29 and e-services,

03:31 present an exciting new frontier

03:35 in the efforts to tackle the challenges of fraud

03:38 and

03:39 corruption.

03:40 These tools

03:42 That is e-government systems and e-services tools

03:47 are closely related to improved outcomes

03:51 in government effectiveness and perceptions of corruption.

03:56 Some

03:57 preliminary analysis by the Govtech team

04:00 show that

04:02 governments that rank high on UN Online Services Index.

04:07 Perform better

04:09 in corruption ranking

04:11 and government effectiveness indicators,

04:14 and this will not be surprising.

04:16 It is really intuitive that

04:19 if

04:20 governments are using

04:22 a lot of these services

04:24 or digital methods to provide services to citizens,

04:27 that will reduce

04:29 interactions between human beings and also reduce

04:33 the opportunities to be able to ask for

04:37 and receive bribes.

04:39 Also,

04:40 if government is using digital tools to provide services,

04:43 I think governments will be able to reach more people

04:47 and we will be able to reach them in a more transparent ways.

04:52 So

04:53 we really think that these tools can be able to help in many ways

04:59 to control fraud and corruption.

05:03 As highlighted in the report that Tracy alluded to,

05:06 which is being discussed this morning.

05:09 There are significant opportunities for

05:12 and benefits from using digital tools to

05:15 tackle fraud and corruption challenges in the public sector.

05:19 Then these are benefits that extend beyond

05:22 just detecting and preventing fraud and corruption.

05:26 In fact,

05:27 studies show that taking advantage of

05:30 the full potential of government digitization

05:33 can free

05:34 up to $1 trillion

05:38 annually in global economic value

05:41 through lowered costs

05:43 and improved operational performance.

05:46 At the same time,

05:48 We also need to know that digital tools are not

05:51 a panacea to all the challenges of fraud and corruption,

05:55 as the country case studies in the report that Tracy mentioned

05:59 illustrate.

06:01 Technology is most effective

06:03 when it is paired with traditional fraud detection

06:07 and prevention methods,

06:09 and

06:09 they are integrated with analog components of the reform.

06:14 I also want to emphasize

06:16 that the rule of law is very fundamental

06:20 to tackling corruption

06:22 and fraud in any

06:25 country context.

06:27 I look forward to a very rich discussion today

06:31 and I want to thank the organizers and also the speakers and the discussions.

06:36 Thank you so much.

06:37 Back to you,

06:38 Tracy.

06:41 Thanks,

06:41 Ed.

06:41 We're um moving straight to Hunt and,

06:44 uh,

06:45 Hunt,

06:45 you have the floor of the,

06:47 uh,

06:47 the presentation as well.

06:48 Thanks.

06:55 Thank you very much,

06:56 uh,

06:57 Tracy and Ed,

06:58 for the marvelous,

06:59 uh,

06:59 introduction.

07:00 I'm going to,

07:02 uh,

07:02 share my screen at the moment.

07:06 Let's see,

07:06 and I will.

07:08 There we go.

07:09 OK.

07:09 Hopefully,

07:10 everybody can see

07:11 uh

07:12 the presentation.

07:14 So,

07:15 today,

07:15 uh,

07:16 the event on finding fraud,

07:17 and this is based,

07:18 uh,

07:19 as Tracy and Ed,

07:20 uh,

07:21 discussed,

07:21 uh,

07:22 on a report called Finding Fraud GovTech

07:25 and fraud detection in Public Administration.

07:28 Um,

07:29 And this really focuses when we talk about

07:32 public administration we're talking about in this report 3

07:36 separate information systems in the public sector that's

07:42 Uh,

07:43 ImusS and HR,

07:44 I will be discussing,

07:45 uh,

07:46 procurement and ITMIS,

07:47 and our colleague Izza will be discussing,

07:49 uh,

07:50 the HR,

07:51 uh,

07:51 systems.

07:52 Um,

07:54 so I am,

07:54 uh,

07:55 a senior procurement specialist in the governance procurement global unit.

07:59 Uh,

07:59 I had the pleasure of TTLing this particular report,

08:03 uh,

08:04 and I look forward to

08:05 going through,

08:07 uh,

08:07 and discussing.

08:08 Uh,

08:09 some of the great findings of this report.

08:11 Um,

08:12 we'll talk about,

08:13 uh,

08:13 you know,

08:14 what the major issues are which Tracy and Ed alluded to,

08:17 uh,

08:18 what challenges exist,

08:20 uh,

08:20 and what,

08:21 uh,

08:21 digital tools can,

08:22 uh,

08:23 can countries deploy to combat

08:25 corrupt practices

08:27 and other drivers that should be considered,

08:29 uh,

08:30 what are some applications and benefits,

08:31 and then key recommendations

08:33 and next steps.

08:35 Um,

08:37 so,

08:37 uh,

08:37 thanks,

08:38 Tracy and Ed for,

08:39 uh,

08:39 in,

08:40 um,

08:41 explaining why this is a major issue.

08:43 One thing I did want to emphasize is that

08:46 fraud and corruption worsens during emergencies,

08:49 uh,

08:49 and the increased demand and time pressures for acquisition of,

08:53 uh,

08:53 remedial goods and services,

08:55 especially

08:56 during COVID-19 has been a major,

08:58 uh,

08:59 issue.

08:59 So,

09:00 and they can lead

09:01 to the relaxation of procurement and inspection.

09:06 Creating an increased risk of selection of unqualified or fictitious suppliers

09:12 and delivery of poor

09:14 quality or non-existence,

09:16 um,

09:17 goods and services.

09:18 So,

09:19 uh,

09:19 uh,

09:20 just wanted to emphasize that in the introduction,

09:23 um.

09:24 And I will move right on to why,

09:26 uh,

09:26 what challenges exist.

09:28 Uh,

09:29 and,

09:29 uh,

09:30 so here we have,

09:31 uh,

09:31 some of the major challenges that do exist and that we do,

09:34 uh,

09:35 we do

09:36 deal with and not,

09:37 uh,

09:37 no technology or digital tools.

09:39 Can guarantee success when it comes to government reform efforts,

09:43 right?

09:44 Um,

09:44 and studies have shown that there,

09:46 uh,

09:47 that there is a skills and resource gap when it comes to technology

09:52 and,

09:53 uh,

09:54 digital.

09:57 In the public sector,

09:59 and political will,

10:00 uh,

10:01 it needs to be installed,

10:03 uh,

10:04 digit,

10:04 uh,

10:04 when you're,

10:05 there needs to be political will,

10:07 uh,

10:07 when,

10:08 uh,

10:08 we have digital,

10:10 uh,

10:10 anti-fraud systems,

10:11 what's one of the issues that we deal with,

10:14 uh,

10:14 and also there needs to be a follow-up on the results

10:18 uh of on with appropriate,

10:20 um,

10:20 sanctions in the absence of many of the countries,

10:24 uh.

10:25 Uh,

10:26 uh,

10:26 where the systems are needed most.

10:28 Also,

10:29 poor infrastructure,

10:31 including,

10:31 uh,

10:32 such as intermittent internet access.

10:35 Uh,

10:36 and lack of computing,

10:38 uh,

10:38 power,

10:39 uh,

10:39 to process the data and applications involved in procurement and IT systems.

10:44 Also,

10:45 poor financial management strategy can result

10:47 in reduced usage and minimal standards

10:51 and corrupt financial,

10:52 uh,

10:53 reporting practices,

10:54 uh,

10:55 and automation of bad practices and out of date institutions

10:59 and even lack of translated materials in many of our,

11:03 uh,

11:03 client countries.

11:05 Um,

11:06 so these are,

11:07 uh,

11:07 uh,

11:08 additionally,

11:09 governments,

11:09 uh,

11:10 can work in developing countries can work in silos,

11:13 and that's one of the issues

11:14 that we deal with,

11:15 uh,

11:16 many times in the topic we're discussing today.

11:18 Many times the public procurement agency,

11:21 uh,

11:21 Doesn't,

11:22 uh,

11:23 um,

11:24 interface well with the,

11:25 um,

11:26 uh,

11:26 with the Ministry of Finance,

11:28 meaning that the information system,

11:30 the e-procurement system for the public procurement agency

11:33 isn't integrated with the data from the IMI

11:35 system and therefore it can be very challenging,

11:38 uh,

11:38 to share information back and forth.

11:41 Um,

11:42 so what are the challenges in procurement?

11:45 Uh,

11:45 combating fraud and corruption in procurement,

11:48 uh,

11:49 is a central

11:51 in digital,

11:52 uh,

11:52 detection.

11:53 Procurement

11:54 is where most fraud,

11:55 uh,

11:56 and corruption cases

11:57 and losses occur,

11:59 uh,

11:59 where governments and where governments spend the most money,

12:03 often,

12:04 uh,

12:04 financed

12:05 by

12:06 international donors.

12:08 Um,

12:08 but few e-procurement systems currently

12:10 include ex-ante fraud detection programs,

12:14 which we refer to as governance or integrity filters

12:18 in the routine purchasing of uh,

12:20 uh,

12:21 in routine purchasing.

12:23 And It appears that there's no such

12:25 programs that monitor large scale Tinder transactions,

12:29 which are serious losses where serious losses are routinely incurred.

12:33 This could be in infrastructure projects

12:36 and in large IT projects.

12:40 And IMIS,

12:41 uh,

12:42 platforms are cited as useful in fraud detection.

12:45 In fact,

12:46 we'll,

12:46 uh,

12:46 demonstrate some very nice,

12:48 um,

12:49 or not demonstrate,

12:50 but we'll,

12:50 uh,

12:51 discuss some,

12:52 uh,

12:53 IMIS modules that can be,

12:55 uh,

12:55 implemented.

12:56 Uh,

12:57 but IMIS in general is expensive,

12:59 complex,

12:59 and difficult to install and operate.

13:02 Um,

13:02 additionally,

13:03 IMIS projects can have,

13:05 uh,

13:05 other issues

13:07 including unsound project design

13:09 and lack of necessary underlying finance reforms.

13:13 Um,

13:14 and

13:15 I,

13:15 uh,

13:16 systems can be vulnerable to,

13:18 uh,

13:18 several fraud and corruption,

13:20 uh,

13:20 schemes such as

13:22 the misallocation of budget items.

13:25 Processing of inflated payments to shell companies or phantom vendors

13:30 and payments to offshore accounts as part of money laundering schemes.

13:35 Um,

13:36 now we'll start talking about some of the digital tools and some of the,

13:40 uh,

13:40 fraud schemes and algorithms.

13:43 Um,

13:45 And in,

13:46 in recent years,

13:47 there has been

13:49 a

13:50 Uh,

13:51 in recent years,

13:51 there have,

13:52 uh,

13:54 been breakthroughs in digital technologies that have expanded

13:57 the scope of reform,

13:59 uh,

13:59 possibilities and provided an array of new,

14:02 uh,

14:02 tools to governments to help

14:05 them improve,

14:06 uh,

14:06 governance outcomes and control corruption.

14:09 Um,

14:10 and so,

14:11 effectively use of digital tools,

14:14 uh,

14:15 requires

14:16 service.

14:18 Let me make sure I'm on the right page here.

14:19 Yup,

14:19 it requires uh services of high qualified,

14:22 uh,

14:23 uh,

14:24 diligent and ethical development professionals,

14:27 a proper training in anti-fraud measures and,

14:30 and equipped with tools to detect and prevent it,

14:33 such as,

14:33 uh,

14:34 the digital fraud tools we'll be discussing.

14:36 And in GovTech,

14:37 uh,

14:38 which

14:38 we have many e-government tools and e-services is

14:41 closely related to improving outcomes and government effectiveness

14:45 and perception of corruption.

14:48 So,

14:49 um,

14:50 today we're gonna be talking about,

14:52 uh,

14:52 we're gonna start talking about some of the detailed

14:55 integrity filters,

14:56 which are sophisticated algorithms

14:59 to detect and prevent fraud and irregularities that can

15:03 be embedded in any procurement and IMI system.

15:06 And these filters can run proactively,

15:09 uh,

15:09 which means,

15:10 uh,

15:10 ex ante,

15:12 uh,

15:12 to identify possible fraud over before,

15:15 uh,

15:15 bids are evaluated or payments are approved or ex post

15:19 against procurement data stored in historical databases.

15:24 Um,

15:27 Uh,

15:28 and here for our integrity filters,

15:30 uh,

15:30 we exploit,

15:31 uh,

15:32 many of the,

15:34 uh,

15:34 electronic procurement,

15:36 uh,

15:36 data that we have.

15:37 So,

15:38 here we can block non-compliant transactions.

15:41 Uh,

15:42 we can provide instant proactive alerts

15:44 of possible fraud,

15:46 uh,

15:47 and more so with electronic procurement,

15:50 we can review 100% of all.

15:53 Transactions were previously

15:54 uh during the paper-based processes you could only do one transaction at a time

15:59 and it allows uh for a real-time remote monitoring by

16:04 donors or oversight agencies if need be and there can be

16:08 uh

16:09 uh create a detailed audit trails and digital.

16:11 Evidence for investigators,

16:14 um,

16:15 uh,

16:15 and we also identify evidence of previous ongoing misconduct in,

16:20 uh,

16:21 in

16:21 historic databases,

16:23 and we have different types

16:25 of

16:26 Uh,

16:27 levels of fraud,

16:28 right?

16:29 Uh,

16:30 and,

16:30 uh,

16:31 in reports we have significant procurement statistics,

16:34 uh,

16:35 where these would be numbers

16:37 of awards to certain contractors by certain approving officials.

16:42 We also have reports on economy and

16:43 efficiency indicators that ensure the selection of the

16:47 best product at the best price,

16:49 um,

16:50 uh,

16:51 compliance reports,

16:52 obviously,

16:53 uh,

16:53 where contractors in violation of procurement rules.

16:55 Can be identified,

16:57 uh,

16:57 maybe a short bid notice could be also identified,

17:01 uh,

17:01 or,

17:02 or bids from a debarred,

17:04 uh,

17:04 company.

17:05 We also have SPQQD reports,

17:08 and these are select,

17:10 uh,

17:10 selection price,

17:12 uh,

17:12 qua quantity,

17:13 quality,

17:14 and delivery indicators that can point to fraud,

17:17 uh,

17:17 waste,

17:17 or abuse.

17:19 Uh,

17:19 and then we also have,

17:20 uh,

17:21 information on collusive bidding and big rigging which we're gonna dive into.

17:25 Um,

17:26 and here are some of the common fraud and corruption schemes,

17:30 uh,

17:30 in procurement and that do reach into

17:33 IFMIS systems,

17:35 uh,

17:35 when we start talking about the purchasing,

17:37 because in public procurement,

17:40 Generally,

17:41 uh,

17:41 we,

17:41 we,

17:42 we,

17:42 the public procurement agency works with

17:45 up until the contract award and then at the contract

17:49 management it's handed off,

17:50 uh,

17:50 to the procurement entity itself,

17:53 uh,

17:53 many times,

17:54 uh,

17:54 to manage that and also deal with

17:57 the invoicing,

17:58 receiving,

17:58 and payments.

17:59 So we'll talk about,

18:00 um,

18:01 A couple tendering frauds,

18:03 uh,

18:04 collusive bidding,

18:05 bid rigging,

18:05 kickbacks,

18:06 conflicts of interest,

18:07 and purchasing frauds,

18:09 um.

18:10 Uh,

18:11 such as,

18:12 uh,

18:12 uh,

18:12 false inflated and duplicate invoices,

18:16 shell companies,

18:16 phantom vendors,

18:17 and purchasing,

18:18 uh,

18:19 purchases for,

18:20 uh,

18:20 use,

18:20 uh,

18:21 or resale.

18:23 Um,

18:24 now we,

18:24 we have,

18:25 we've color coded the indicators,

18:27 uh,

18:27 to help get an idea of where they,

18:29 they can be found and how they can be used.

18:32 So the reds are,

18:33 uh,

18:34 real-time blocks or alerts for significant indicators.

18:37 The browns

18:38 are preprogrammed reports.

18:39 Reports,

18:40 uh,

18:40 for other common procurement fraud schemes or waste or abuse.

18:44 The orange are other less common reports to be listed

18:47 on the handbook or online,

18:49 and blue are links to public,

18:51 uh,

18:51 procurement

18:52 records,

18:52 telephones,

18:53 and address information,

18:54 which is very helpful.

18:56 So let's talk about the first,

18:58 um.

18:59 Uh,

19:00 the first report,

19:01 uh,

19:01 which we have,

19:02 or the first scheme,

19:03 which is bid rigging.

19:05 Uh,

19:06 bid rigging refers to the secret agreements between favored bidder

19:11 and procurement personnel,

19:13 uh,

19:14 to improperly manipulate the procurement process to steer contract award

19:18 to a favored bidder and exclude other bidders,

19:21 often as a result of corruption.

19:24 So there are many different common bid

19:27 rigging schemes including change order abuse,

19:30 uh,

19:31 excluding qualified bidders,

19:33 uh,

19:33 leaking the bid information beforehand.

19:36 The manipulation of bids,

19:38 um,

19:39 rigged or targeted specifications happens quite a bit,

19:43 splitting up purchases,

19:44 and

19:45 you can,

19:46 uh,

19:46 identify these by taking a look at the sample indicators that we've identified,

19:51 and these are all in the report,

19:53 uh,

19:54 and these are the primary data sources here that you would find.

19:58 Um,

19:59 each of these,

20:00 uh,

20:01 uh,

20:01 of these various indicators,

20:03 and algorithms can be created,

20:05 uh,

20:06 either through reports or BI or ex ante red

20:09 flag indicators into your e-procurement system to identify,

20:13 uh,

20:13 this.

20:14 The next one is collusive bidding,

20:17 uh,

20:17 and collusive bidding refers to agreements,

20:20 uh,

20:20 by contractors or suppliers to,

20:23 uh,

20:24 cooperate

20:25 in the

20:25 bidding process in order to avoid competition

20:28 and inflate prices to artificially high levels.

20:32 It can occur in small and large contracts

20:36 where collusive bidding is well established,

20:38 uh,

20:39 prices can rise

20:41 significantly.

20:42 It's sometimes over 100%.

20:44 Um,

20:45 and there are various types of,

20:47 uh,

20:48 of this collusive bidding.

20:50 One is complementary bids.

20:51 These are,

20:52 uh,

20:52 can also be known as protective courtesy or shadow bids,

20:57 um,

20:57 uh,

20:58 that are intended to merely give an,

20:59 uh,

21:00 appearance

21:01 of a genuine bid,

21:02 but it's not,

21:02 uh,

21:03 they're not,

21:04 uh,

21:04 meant to secure the buyer's acceptance or to win.

21:08 There's also bid rotation,

21:09 uh,

21:10 where participants in bidding schemes or ongoing cartel activity

21:15 often rotate winning bids on geographical

21:18 areas or based on job type or uh

21:21 timing.

21:22 That's,

21:22 that happens quite a bit.

21:24 And then,

21:25 uh,

21:25 bid suppression where,

21:26 uh,

21:27 big breeding streams to succeed groups,

21:30 members must prevent outside companies from bidding,

21:33 and this can be done by either a payoff or,

21:36 you know,

21:36 through threats or potentially violence.

21:38 And then there's market division,

21:40 uh,

21:40 the cooperating companies may divide markets,

21:43 uh,

21:43 and product lines and agree to not compete on each other's territory.

21:48 So that

21:49 all these happen quite a bit,

21:50 and here we have simple indicators that can illustrate.

21:53 Um,

21:55 uh,

21:56 it,

21:56 it,

21:56 by,

21:57 um,

21:58 measuring these,

22:00 uh,

22:00 these can provide,

22:01 uh,

22:01 help

22:02 for risk,

22:03 and then you can find these type of indicators

22:05 in,

22:06 uh,

22:06 e-procurement system and IMIS systems,

22:09 uh,

22:10 under,

22:10 uh,

22:11 these types of data sources.

22:13 And let's talk about,

22:14 uh,

22:14 our first,

22:15 uh,

22:16 purchasing fraud,

22:17 uh,

22:17 and,

22:19 uh.

22:24 So a contractor or supplier can.

22:31 This inflated intent to defraud either acting alone or in collusion with a con.

22:38 Uh contracting or personnel,

22:39 um,

22:40 and so you have,

22:41 uh,

22:41 false invoices,

22:42 uh,

22:43 which,

22:44 um,

22:45 our invoice information does not match the purchasing order

22:48 receiving or payment information.

22:51 Um,

22:52 also there can be,

22:52 um,

22:52 uh,

22:52 sequential,

22:56 uh,

22:56 invoice,

22:57 uh,

22:57 numbers and then inflated invoices where the invoice

23:01 price.

23:02 Quantities are greater than the purchasing order price,

23:06 and the total payments are greater than the total invoice amounts

23:10 and then duplicate invoices,

23:12 uh,

23:12 where the invoices with the same numbers,

23:14 dates,

23:14 and amounts,

23:15 uh,

23:16 are paid.

23:16 So that's,

23:17 uh,

23:17 these are,

23:18 uh,

23:19 the different indicators that we can use,

23:21 uh,

23:22 to be able to identify,

23:23 uh,

23:24 this type of scheme.

23:25 Um,

23:26 additionally,

23:27 we have shell companies.

23:28 And with shell companies,

23:30 uh,

23:31 these are vendors referred to firms that are secretly owned by,

23:35 uh,

23:35 procurement or agency officials employed by the procurement,

23:38 uh,

23:39 agency.

23:40 Such schemes,

23:42 uh,

23:42 such schemes,

23:43 uh,

23:43 are typically classified as conflict of interest

23:46 under the general category of corruption.

23:49 Um,

23:51 And shell companies can also refer to fictitious companies set up

23:55 by corrupt officials

23:57 uh to act

23:58 and portray suppliers and subcontractors in order to receive bribes

24:03 and hidden assets.

24:04 And here we can see the various indicators,

24:07 uh,

24:08 that you could identify,

24:10 uh,

24:10 uh,

24:11 from these,

24:11 uh,

24:12 sources and include those in your algorithms,

24:15 uh,

24:16 for identifying shell companies.

24:18 And then we also have,

24:20 uh,

24:20 phantom.

24:20 Vendors and Phantom vendors are,

24:23 uh,

24:25 are our procurement

24:26 or payment personnel

24:28 that can create a fictitious,

24:29 uh,

24:30 contractor,

24:31 consultant,

24:32 or vendor

24:33 or supplier that does not provide any goods

24:35 or services in order to embezzle project funds.

24:39 These fictitious companies often provide consulting.

24:43 And other hard to verify services such as a repair work

24:47 or deliver consumables rather than tangible goods

24:51 and services that can be uh later verified.

24:54 Um,

24:54 and then we also have the resale,

24:57 uh,

24:57 or purchase of personal use and resale,

25:00 um,

25:01 or uh diversion.

25:03 Um.

25:05 Some indicate sample indicate

25:07 here,

25:07 uh,

25:08 uh,

25:08 you can have a,

25:09 a,

25:10 a,

25:10 a different ship to address,

25:12 a high number of purchases of,

25:14 uh,

25:14 certain items susceptible to personal use such as fuel or laptops,

25:19 uh,

25:19 or gas,

25:20 uh,

25:21 unexplained,

25:22 uh,

25:22 spike in purchase of such items,

25:25 and you can get this information,

25:26 uh,

25:27 from these data sources here from the the vendor and product codes,

25:31 PO invoicing,

25:32 shipping information records for.

25:34 Um,

25:36 So,

25:37 uh,

25:37 a lot of these also have,

25:39 uh,

25:39 to,

25:40 uh,

25:40 one way of,

25:41 uh,

25:42 not validating,

25:43 but identifying risks is by applying Benford's law,

25:48 which allows you to,

25:49 um,

25:51 to take a,

25:53 uh,

25:54 Benford's law works with,

25:55 um,

25:56 any set of numbers generally.

25:58 Uh,

25:59 take the,

25:59 the populations of all of the countries in the world,

26:02 if you put those on a piece of paper,

26:04 you would find that,

26:06 um,

26:07 That the number 1,

26:09 would be there about 30% of the time,

26:11 uh,

26:12 the number 2 would be there about 18% of the time,

26:14 and the number 3,

26:16 it kind of goes down in a logarithmic fashion.

26:19 It's uh

26:20 an interesting,

26:21 um.

26:22 Uh,

26:23 uh,

26:23 fraud detection,

26:24 uh,

26:25 resource that can be used,

26:26 obviously,

26:27 uh,

26:28 and it can help to indicate fabricated numbers,

26:32 uh,

26:33 and fraud.

26:33 And,

26:34 uh,

26:35 again,

26:35 uh,

26:36 you would need to validate this,

26:37 uh,

26:38 before you,

26:39 uh,

26:39 move forward with it,

26:40 but it's a definitely

26:41 an,

26:42 a way of identifying,

26:44 uh,

26:44 a risk.

26:46 Um,

26:46 there are many,

26:47 uh,

26:48 commercial products on,

26:50 uh,

26:51 the market for

26:52 detection of ifus,

26:54 where in

26:55 e-procurement,

26:56 uh,

26:56 we don't have,

26:57 uh,

26:57 there,

26:58 there are some,

26:59 um,

27:00 but they don't,

27:01 um,

27:02 they aren't fully integrated in with e-procurement systems as of yet.

27:07 Um,

27:08 and they,

27:09 lots of times they only go into the contract award.

27:12 Um,

27:13 whereas for,

27:15 uh,

27:15 IMIS systems there are,

27:17 uh,

27:17 modules

27:18 on the market that you can,

27:20 uh,

27:21 purchase and install.

27:22 Uh,

27:23 in fact,

27:23 um,

27:25 So,

27:26 uh,

27:26 there are a number of robust,

27:28 uh,

27:28 commercial fraud detection prevention systems that can be installed or linked,

27:32 uh,

27:32 to IFMIS and ERP such as SAP,

27:35 um,

27:36 and these systems can provide continuous monitoring and ex-ante alerts,

27:42 uh,

27:42 or provide,

27:43 uh,

27:43 fraud,

27:44 uh,

27:44 uh,

27:45 of potential fraud,

27:46 many related to accounts payable transactions.

27:50 Um,

27:50 so here we have,

27:52 uh,

27:52 the SAP HANA fraud detection and you can see many of the features that this provides.

27:58 Um,

27:59 more than half of ITMIS systems,

28:01 uh,

28:01 funded by the World Bank or in,

28:04 uh,

28:04 in developing countries

28:05 do use these,

28:06 uh,

28:07 um,

28:08 commercial products,

28:09 uh,

28:10 and then,

28:11 then others can be,

28:12 uh,

28:13 homegrown.

28:14 Um,

28:15 this particular SAPPANA fraud management.

28:20 I programmed to detect and investigate and

28:22 prevent fraud in day to day processes including

28:26 uh order cash,

28:28 uh,

28:28 procure to pay and product,

28:30 uh,

28:30 and planned a product,

28:32 request a service,

28:33 and core capabilities.

28:35 Another

28:36 product out on the market uh is Appen,

28:39 and it's an AI

28:41 learning,

28:42 uh,

28:43 a machine learning audit and spend management application that links to IFMMIS

28:48 and other business management systems.

28:51 Um,

28:51 it's similar to the SAP module,

28:53 uh,

28:53 and it can identify anomalies

28:56 and possible fraud in accounts payable.

28:59 Um,

28:59 one interesting thing with apps in is that you have,

29:02 uh,

29:03 they have,

29:04 uh,

29:04 the ability to use AI technology to read and interpret written content in documents

29:10 to identify anomalies and compliance issues.

29:13 Um,

29:15 and,

29:15 uh,

29:16 and it can be used to read the text

29:19 of receipts,

29:20 uh,

29:20 and disallowed items.

29:22 We also have Oversight Systems has a lot,

29:25 and,

29:25 uh,

29:26 Galvanize has,

29:27 uh,

29:28 ACL,

29:29 uh,

29:29 Essentials is another product,

29:31 uh,

29:31 on business.

29:33 Um,

29:33 what other drivers could be,

29:34 should be considered?

29:36 Well,

29:36 other drivers,

29:37 uh,

29:38 is we need to make sure we have the data

29:40 available,

29:41 uh,

29:41 so that,

29:42 uh,

29:42 it can be used.

29:43 The data needs to be available from.

29:46 Um,

29:48 uh,

29:48 from the IMMIS and,

29:50 uh,

29:50 procurement systems,

29:52 uh,

29:52 and these also,

29:53 there should be data privacy,

29:55 uh,

29:56 and protections around this data so that the folks that

30:00 are only allowed to view the data can view it.

30:03 Um,

30:03 we also need to make sure that the,

30:06 the systems are interoperable,

30:08 uh,

30:08 and that,

30:09 uh,

30:10 we're integrating with,

30:11 uh,

30:12 existing,

30:13 uh,

30:13 detect.

30:17 Um,

30:18 so we will,

30:19 I think the big thing that we found with the study is that,

30:23 uh,

30:36 This

30:46 Hunt we're having difficulty hearing you.

30:50 And you have,

30:51 you,

30:51 you are breaking in and out.

30:56 Let me see if I,

30:57 OK,

30:57 is that any better right there?

31:02 Mhm.

31:04 A little better.

31:05 Yes,

31:05 OK,

31:07 OK,

31:08 very good.

31:08 Um,

31:09 let me

31:14 Let me get to the right page here.

31:19 OK,

31:19 what are,

31:20 uh

31:22 So

31:24 Uh,

31:24 application benefits.

31:25 Uh,

31:25 detecting fraud and corruption,

31:27 particularly in procurement,

31:28 is a major benefit.

31:29 Most current procurement,

31:31 uh,

31:31 fraud detection algorithms are expost,

31:34 and we would prefer these to be ex ante.

31:36 Uh,

31:37 and e- procurement systems can certainly help with that.

31:41 Um,

31:42 for,

31:42 uh,

31:43 anti-fraud systems that can be installed or linked to commercial

31:46 ITMMIS systems to improve continuous monitoring and ex-ante alerts,

31:51 um,

31:52 and similar,

31:53 uh,

31:53 functions can be programmed in homegrown systems.

31:57 So our key recommendations that we have

32:00 is

32:01 we want to ensure,

32:03 um,

32:04 the importance of moving to ex ante fraud detection.

32:07 Uh,

32:08 we also need,

32:09 uh,

32:10 need

32:11 to ensure that we're using a stronger fraud detection algorithms,

32:15 uh,

32:15 that,

32:16 uh,

32:17 traverse multiple systems such as e-procurement and IFMIS.

32:21 Um,

32:21 and the,

32:22 uh,

32:23 fraud detection algorithm should be tailored to count,

32:26 uh,

32:26 to countries where they are installed.

32:29 Um,

32:29 the fraud detection technology is most effective when it's

32:32 integrated in with traditional detection and prevention methods.

32:36 Um,

32:36 and,

32:37 and should,

32:38 uh,

32:38 the automated fraud detection should be

32:40 extended to project implementation stage.

32:43 Um,

32:44 And I do want to discuss real quickly

32:47 what the next steps,

32:48 what we think would be uh nice to do is

32:51 create a prototype for ex ante digital fraud detection program for procurement

32:57 and put and installing fraud detection.

33:00 Systems to run remotely and independent oversight organizations.

33:04 Also,

33:04 I think it's very important to include fraud detection systems into gov tech

33:09 projects where there is political will

33:11 and uh digital infrastructure capacity and,

33:14 uh,

33:15 enthusiasm.

33:16 So thank you very much.

33:17 I,

33:18 I look forward to IS's presentation.

33:20 All the best.

33:22 Thanks,

33:22 Han.

33:23 Let's move uh swiftly to,

33:25 to ISA and to hear about the fraud detection in

33:28 HR systems.

33:29 Over to you,

33:30 Isa.

33:31 Thank you so much,

33:32 Tracy and Han.

33:33 So let me,

33:34 uh,

33:34 connect my screen.

33:40 And

33:40 so I hope that is

33:42 working and you can see it right now.

33:46 Um,

33:48 OK.

33:49 So,

33:50 let me begin.

33:50 Thank you so much,

33:51 Han,

33:52 for your presentation on the IFMIS system,

33:55 um,

33:57 And thank you so much,

33:58 Ed and Tracy,

33:59 for your initial comments as well.

34:01 Um,

34:02 I wanted to

34:04 start off a little bit by talking about,

34:07 um,

34:08 you know,

34:08 corruption,

34:08 good governance,

34:09 digital technology,

34:10 sort of setting the stage,

34:12 then take you to a more focused discussion on HR

34:15 systems and how we can apply digital tools to HR systems

34:19 to help detect and prevent fraud and corruption.

34:22 And then I also want to take some time towards the

34:24 end of my presentation to talk about the analog components.

34:27 Um,

34:27 Hunt has mentioned them in.

34:29 Very effective fashion already.

34:31 Um,

34:31 so it doesn't bear repeating,

34:33 but I will highlight some of the

34:35 particular analog compliments as they relate to HR systems.

34:40 Um,

34:40 so,

34:41 moving on to the first section,

34:42 which is kind of looking at corruption,

34:44 good governance,

34:47 And digital technologies.

34:51 Um,

34:51 can,

34:51 can you show us the presentation?

34:53 We cannot

34:54 see from our side.

34:55 You can see it,

34:56 my apologies.

34:57 Let me

35:01 Let me do it again.

35:09 Please let me know if this works and apologies for the

35:13 hiccup.

35:14 Does this work now?

35:17 Yes,

35:17 can you make the screen?

35:19 Yes,

35:19 OK,

35:20 it works well.

35:20 There we go.

35:21 OK.

35:22 Month 11 of,

35:24 um,

35:25 quarantine and we still haven't figured it out apparently.

35:27 OK,

35:28 I think so now with the hiccups past us,

35:30 let me begin again.

35:32 Um,

35:32 so,

35:32 as I said,

35:33 this part of the presentation will focus on HR systems specifically,

35:37 and how we can use digital tools to help

35:39 fight fraud and corruption in the public sector.

35:42 And towards the end,

35:43 I would also reflect a little bit on the analog complements to digital reform.

35:48 So,

35:48 kind of setting the stage,

35:49 um,

35:50 I think Edge and Tracy in their comments

35:52 really talked about the importance of digital tools

35:55 to fight corruption.

35:57 Uh,

35:57 we find that there is a very intrinsic link between

36:01 Um,

36:01 you know,

36:01 corruption perceptions and the ideas that we have regarding good governance,

36:06 these are measured very effectively in global indices as well.

36:10 So,

36:10 as you see on the charts over here,

36:12 on the left and the right,

36:14 on the left,

36:14 we have the worldwide governance indicators and sort

36:17 of these pillars that we consider to be,

36:20 um,

36:20 you know,

36:20 Good governance or ideas like rule of law,

36:23 regulatory quality,

36:24 voice and accountability,

36:26 govern government effectiveness.

36:28 Um,

36:28 and when we find,

36:29 uh,

36:29 what we find in our data analysis,

36:32 and this is taken from a very good publication that was produced by the World Bank,

36:35 um,

36:36 is that control of corruption is very closely tied to a lot of other.

36:41 Um,

36:41 indices related to good governance.

36:43 So,

36:43 a country's performance on many elements of good governance

36:47 are closely tied towards its performance on the control of corruption.

36:51 And the control of corruption pillar is also very closely tied with how citizens

36:56 perceive their government and how much trust they place in their political actors.

37:00 And that's what you're seeing on the chart on the right-hand side.

37:03 Over here.

37:04 Um,

37:05 what it shows essentially,

37:06 if you look at sort of the,

37:07 the,

37:08 the scale here,

37:09 as you move from 0 to 100 in terms of perceptions of corruption,

37:13 100 indicates that,

37:15 uh,

37:15 people perceive corruption,

37:16 the government to be absolutely clean.

37:18 So,

37:18 there are no instances of corruption

37:20 that are taking place.

37:22 And on the left hand side,

37:23 what you're seeing is that as you go up from 0 to 7,

37:27 Has greater public trust in politicians.

37:29 And here,

37:30 although this is not a causal link,

37:32 it,

37:32 it does very much show that there is a close correlation between these two elements,

37:36 which is to say that as this perception of corruption,

37:40 um,

37:40 decreases,

37:41 as

37:42 more and more people think that government is clean,

37:44 that officials are not corrupt,

37:46 they are.

37:46 And greater and greater trust in public officials as well.

37:50 So,

37:50 it's very important for governance,

37:52 um,

37:53 measurements and governance outcomes.

37:55 At the same time,

37:56 what we see is that there hasn't been much improvement,

37:59 right?

37:59 So,

37:59 if you look at,

38:00 again,

38:00 the WGI's

38:02 index on control of corruption,

38:03 and you compare what's been going on over the past decades,

38:07 You see here for the majority of the regions,

38:09 there hasn't been much progress.

38:10 In fact,

38:11 percentile ranking has actually declined in regions such as LAC,

38:15 Middle East,

38:16 South Asia,

38:17 Sub-Saharan Africa,

38:18 even in reporting that we see in

38:20 other measures such as Transparency International's,

38:23 uh,

38:23 Corruption Perceptions Index,

38:25 what you find,

38:26 um,

38:27 is that

38:28 a significant number of countries

38:30 have been performing very poorly year on year when it comes to corruption outcomes.

38:36 Um,

38:37 so with that,

38:38 what are the opportunities that digital technologies can present,

38:41 right?

38:42 Um,

38:42 here in this slide,

38:43 what we're trying to demonstrate and what we were able to find.

38:47 In the course of working on this report,

38:49 is that um a lot of digital technologies are creating

38:52 new frontiers for the fight against fraud and corruption.

38:55 So,

38:56 you see some examples here of governments

38:58 taking direct interventions to fight corruption,

39:01 like in Albania,

39:03 where the government created an SMS-based platform for citizens.

39:07 To be able to report,

39:08 uh,

39:08 when they experience somebody asking for a bribe from

39:11 them in order to access a government service,

39:14 or when they witness somebody else giving a bribe or being asked to take a bribe.

39:18 And this really helped elevate the government's fight against corruption.

39:22 They were able to

39:23 increase the number of corruption investigations that they were taking on.

39:27 And led to positive outcomes.

39:29 Um,

39:29 on the other hand,

39:30 and this is something that Ed alluded to towards the beginning of our discussion,

39:34 is that there are a lot of indirect benefits to digitization as well.

39:37 So,

39:38 as governments,

39:39 uh,

39:39 take on this digitization effort,

39:41 as they digitize more and more government functions,

39:44 what you find.

39:45 Happening is that the opportunities to engage

39:48 in corrupt behavior and fraudulent activities actually

39:51 closes down.

39:51 And so there's this indirect kind of an externality,

39:55 um,

39:56 which is very positive in that there is less corruption taking place.

39:59 And that's the example that we see here,

40:01 um,

40:02 in Afghanistan,

40:03 where,

40:04 The creation of this M-Pesa mobile payment system

40:08 allowed the government to stem a lot of this stiffening

40:11 that was taking place in terms of salary payments.

40:13 Um,

40:13 it was actually very interesting.

40:15 Employees thought that they were receiving a bonus.

40:18 Their incomes were going up by an average of 10%,

40:21 when in fact what was happening was that that 10% was being lost,

40:25 um,

40:25 in the process of them receiving their cash income

40:28 because it was being taken up by either middlemen or ghost workers,

40:31 etc.

40:32 So there was this overall benefit,

40:34 and we're able to capture it in data as well,

40:37 which is what you can see on the right-hand side over here,

40:40 which is,

40:40 again,

40:41 not a causal link,

40:42 but a very close correlation,

40:43 which shows that as you

40:45 um promote,

40:46 as you expand more and more e-services,

40:49 more and more e-Gov functions,

40:51 uh,

40:52 And perceptions of government actually start to decline.

40:55 So again,

40:55 the scale on the bottom,

40:56 0 to 100,

40:57 with 100 indicating clean government,

41:00 and on the left,

41:00 0 to 11 indicating,

41:03 um,

41:03 you know,

41:04 this full development of e-services according to this UN index,

41:08 and this very close,

41:09 tight relationship between the two taking place.

41:12 So,

41:13 moving to sort of the public sector then,

41:15 I mean,

41:15 we've talked about how anti-corruption efforts are

41:18 very important for good governance outcomes.

41:20 We've talked about how digit technologies offer

41:23 a very exciting new frontier.

41:25 But what we find is taking place within the public sector

41:28 is that there's still a lot more room for improvement.

41:31 Of course,

41:32 this also means that every year there's new and greater innovation.

41:36 So,

41:36 every year,

41:36 there is more and more that governments can do.

41:39 Um,

41:39 analysis that we undertook,

41:41 however,

41:41 showed that within the public sector,

41:43 there is still an increasing reliance on very

41:46 traditional mechanisms such as sort of audits,

41:49 um,

41:49 relying on tips from whistleblowers

41:52 to tackle fraud and corruption.

41:54 And we think that much more can be done.

41:56 And in fact,

41:57 when you look at

41:58 the application of digital tools,

42:00 what you find is,

42:01 um,

42:02 that you can run a lot of algorithms,

42:04 you can create a lot of programs that can

42:06 be much more effective in identifying patterns of corruption,

42:09 in identifying risks before.

42:12 And actual fraudulent activity occurs.

42:15 So that's sort of gonna be the focus of my presentation,

42:18 that's the focus of our presentation as well.

42:20 Hunt talked about IFMAIS systems,

42:22 procurement systems.

42:24 Um,

42:24 I wanted to give you kind of an overview of HR systems and how they can be deployed.

42:31 A very important point to make here

42:33 is that whereas traditionally IMIS systems and procurement

42:36 systems have been used in the past,

42:39 um,

42:40 for in terms of government efforts to identify fraud and corruption,

42:44 HR systems have not been similarly utilized.

42:46 Now,

42:47 this makes a little sense because when you're looking at large procurement.

42:51 contracts,

42:51 when you're looking at IMS systems that are tied to government's budgeting,

42:55 uh,

42:55 there's this dollar value that you can assign,

42:57 right?

42:58 So there's a dollar value that you can assign to a fraudulent activity,

43:01 to an instance of corruption,

43:03 which are,

43:04 whereas HR might not present

43:06 a similar cost to the government.

43:08 But the point that we make in our report

43:11 is that any

43:13 instance of somebody breaking the rules when it comes to HR management,

43:17 be it attendance,

43:18 performance,

43:19 etc.

43:20 Uh,

43:20 it represents a loss to the government.

43:23 And so that's something that we're going to talk about in the examples that we share,

43:27 uh,

43:27 which is either the loss in terms of money,

43:29 whereas that you're paying a wage to an employee

43:32 who's not performing,

43:33 so it's a loss of money,

43:35 but also the act of that person.

43:36 not doing their job effectively represents a loss of performance as well,

43:40 a loss of government outcomes.

43:43 Um,

43:43 at the end of the day,

43:44 the core role that the government performs in any society

43:47 is to provide services to citizens,

43:49 is to generate some kind of outcomes at the end of the day.

43:53 So,

43:53 um,

43:54 before I jump into sort of very specific examples and share them with you,

43:58 I wanted to just provide a very quick overview of,

44:01 um,

44:01 an HR system and what it typically looks like.

44:04 In any country.

44:06 Um,

44:06 now,

44:06 this represents sort of the ideal,

44:08 it can vary a lot country to country,

44:10 but,

44:10 um,

44:11 a lot of countries do have some kind of HR system,

44:14 right?

44:14 Um,

44:15 and this HR system will typically capture information

44:18 regarding the public sector employees' biographical information,

44:22 their attendance records.

44:23 Records,

44:24 information about,

44:25 you know,

44:25 the job position,

44:26 title,

44:27 what department they're working in,

44:29 um,

44:29 how much salary they're getting

44:31 in

44:31 form of basic pay or allowances,

44:34 etc.

44:35 and typically these kinds of HR systems are also linked to

44:38 the government's budgeting system such as IFMIS or a procurement system,

44:42 so that you can really use these systems to have an overall view

44:46 of government as well as use it for policymaking and public sector management,

44:50 um.

44:52 So,

44:52 the right-hand side is just a snapshot of South Korea's HR system.

44:56 It kind of represents the idea.

44:58 Um,

44:59 it's meant to just give you an overview of

45:01 the different elements that such a system can include,

45:04 um,

45:05 as I mentioned,

45:05 sort of biographical information,

45:07 attendance records,

45:08 it could be linked to performance evaluation,

45:10 training,

45:11 development,

45:11 etc.

45:12 as well.

45:13 So let me jump into examples.

45:15 What I'm going to do is give you three specific examples of um

45:19 the kind of fraudulent activities that can take place in the public sector

45:24 that are linked to HR

45:26 and how an HR system can effectively be used to tackle them.

45:30 The first example is related to employment and attendance records.

45:35 The example,

45:36 um,

45:37 the examples that I'm going to share here with

45:39 you kind of illustrate what the core challenge is,

45:41 right?

45:41 And the core challenge that we often see in public sectors around the world,

45:45 is that you might have a lot of ghost workers,

45:48 so employees that don't exist,

45:50 in fact,

45:50 but are actually on the government's role.

45:53 They're collecting salaries,

45:54 but they're not actually showing up to do any work.

45:57 And then there's also employees who do exist,

46:00 um,

46:00 but they tend to not show up to work at all for long periods of time,

46:05 with unexcused absences,

46:06 etc.

46:07 An example from India is just one such illustration,

46:10 um,

46:10 where the study from 2016 showed that on any given day,

46:15 Up to 25% of government teachers and up to 33% of healthcare workers

46:20 did not show up to work without a legitimate reason.

46:22 So it's a very significant problem,

46:24 not small at all.

46:26 It does represent a very big cost in terms of wages that are being paid,

46:30 in terms of work effort that's not being undertaken.

46:33 Um,

46:33 and there's several digital solutions that can be applied here.

46:37 The first that we've shared here is,

46:39 is something that's already in use.

46:41 So digital ID systems,

46:43 a lot of us in the audience might be very familiar with this.

46:47 Um,

46:47 their use is expanding across the world.

46:50 The example here that we've shown is from New Guinea

46:53 where they were able to create a biometric ID system for

46:57 employees.

46:58 It was linked to the HRMIS

47:00 and in the process of kind of

47:02 ensuring through biometric registration IDs of employees,

47:06 they were able to eliminate a lot of fictitious positions and ended up saving,

47:10 um,

47:10 Up to $1.7 million US dollars,

47:12 which is a very significant amount.

47:14 The other example which we're also seeing in

47:17 more and more countries around the world,

47:19 is using biometric scanners.

47:21 Um,

47:22 so,

47:22 you know,

47:22 an employee shows up to work,

47:23 they scan their thumbprint,

47:25 they go into the office,

47:27 they do their work,

47:27 they leave at the end of the day,

47:28 they scan their thumbprint again.

47:31 It might not be extremely effective because somebody could show

47:33 up to work and just sit and not do anything,

47:36 but it is one way to ensure that attendance is taking place,

47:39 right?

47:40 And so what you can do is track those attendance records,

47:43 link them to the HRMS,

47:45 make sure that you have um an algorithm that's doing iterative tests,

47:50 and it can raise a red flag when somebody is chronically absent,

47:53 not showing,

47:53 Up to regular working hours,

47:54 etc.

47:55 So,

47:56 this is an example where you're seeing a

47:58 lot of these biometric scanners in government buildings,

48:01 uh,

48:01 but perhaps they're not being utilized most effectively

48:04 because they're not being linked to the HRMS

48:06 and not being run through these kinds of tests,

48:08 which will be very helpful in actually tracking

48:11 where there is a lack of apps,

48:13 a lack of attendance by government employees.

48:16 The second example that I wanted to share with you.

48:19 Um,

48:19 is in the case of recruitment and promotion decisions.

48:22 So,

48:23 here again,

48:23 a very common public sector challenge

48:26 is that,

48:26 um,

48:27 a candidate might submit false records,

48:29 um,

48:30 in terms of,

48:30 you know,

48:31 for the hiring,

48:32 um,

48:32 as well on the other side,

48:34 you might have hiring managers

48:36 who tend to favor certain candidates because they're,

48:40 You know,

48:40 maybe

48:41 tied to them through

48:42 family links,

48:43 or maybe they're just expressing some kind of favoritism

48:46 and impart impartiality to one candidate over the other.

48:50 Um,

48:51 and this can be a very significant challenge as well because at the end of the day,

48:55 what you're taught,

48:56 what you want to promote in a public sector

48:58 are meritocratic principles,

48:59 right?

49:00 So here there are a lot of digital solutions that can be applied to

49:04 the HR system that you have

49:06 with regards to remote recruitment and promotion decisions,

49:09 um,

49:10 to make sure that this kind of bias and this kind

49:12 of sort of fraud and colluding does not take place.

49:15 Um,

49:15 some examples that we've shared here in the report

49:18 and that I wanted to highlight here as well,

49:21 um,

49:21 are things like,

49:22 you know,

49:22 connecting the,

49:24 um,

49:24 HR registry with common registries like academic.

49:28 Registries,

49:28 police reports,

49:29 etc.

49:30 and you can run these tests,

49:31 tests and flag where,

49:33 um,

49:34 you know,

49:34 a candidate that you're hiring actually has,

49:36 you know,

49:37 you can verify their academic credentials,

49:39 you can do a background check and see if they

49:41 have any criminal history that might pose a problem for their

49:45 employment status.

49:46 Um,

49:47 similarly,

49:47 if you're trying to hire somebody who's an internal candidate,

49:51 you can run these kinds of checks in their HR records to show that.

49:56 In the position that they're potentially going to be hired to,

49:59 um,

49:59 you know,

49:59 they have the necessary qualifications,

50:01 the years of experience,

50:02 the academic records,

50:03 the sort of professional skill set,

50:05 etc.

50:06 Uh,

50:07 similar cross-referencing of data elements that you can undertake

50:11 is to flag if one single individual has been responsible for hiring a,

50:15 a,

50:16 a very significant number of people,

50:18 right?

50:18 So,

50:19 if somebody is working at some ministry,

50:23 uh,

50:23 in local government,

50:24 and they're just hiring a lot of people to favoritism,

50:27 to nepotism,

50:29 um,

50:29 if a lot of that sort of is originating from some,

50:32 a small subset of individuals,

50:34 this is a common test that you can apply to kind of flag and explore further,

50:39 whether this is problematic.

50:41 Or if this is just in the course of significant hiring that needs to take place

50:46 and the individual is just performing their duty.

50:48 Um,

50:49 another common example,

50:50 uh,

50:50 that can be applied is that you could have,

50:53 um,

50:53 an algorithm that kind of

50:55 tests for very common last names and family names.

50:58 Um,

50:59 of course,

50:59 this also requires some kind of human oversight

51:02 because some family names are very common,

51:04 uh,

51:04 but it is a potential.

51:05 link between uh families that could indicate somebody's

51:09 hiring because they are related to the candidate.

51:12 Um,

51:12 and so the system could flag these and then,

51:15 um,

51:15 you know,

51:16 actual human auditors could come in and verify

51:19 what is going on in the situation.

51:21 So there's lots of different ways in which really you can apply,

51:25 essentially kind of an add-on to your existing HR system.

51:29 That helps run these tests,

51:30 helps raises these red flags

51:33 of potential problems with regards to

51:35 recruitment decisions and promotion decisions,

51:37 and then you can use human oversight to actually um investigate what's going on

51:42 and stem a problem that might be occurring.

51:44 The third and final example that I wanted to share with you,

51:48 um,

51:48 as an illustration of how HR systems can be used to detect fraud and corruption.

51:53 As in the case of employees' salaries and income.

51:56 And so now here,

51:57 a very big challenge that you see in public sectors around the world,

52:01 um,

52:01 is that you have government officials who take bribes,

52:04 right?

52:05 So this is a very common problem reported by citizens all over the world,

52:08 in order to access a very basic service like

52:11 um getting electricity,

52:13 uh,

52:13 being able to see a doctor,

52:15 get medicine,

52:16 or enroll their child into school,

52:18 they're asked for a bribe by a government worker,

52:21 uh.

52:22 Similarly,

52:22 you have government workers who might be double dipping,

52:24 so this is the idea that an employee is actually,

52:28 um,

52:28 pocketing an allowance,

52:30 an additional benefit,

52:31 an initial bonus,

52:32 or more,

52:33 a double salary,

52:34 which you're not actually eligible for,

52:37 um,

52:37 and typically what's being done right now is that you'll have some kind of auditing,

52:41 some traditional audits to try to verify the salary payments.

52:45 Um,

52:45 but again,

52:46 this is also an avenue where you could use digital tools instead,

52:49 um.

52:50 To kind of stem this problem before it occurs.

52:54 So some of the things that you could do

52:56 is to try again to cross-reference the different data registries and you could say,

53:00 well,

53:01 this individual,

53:02 you could look at their bank transactions,

53:04 their tax filings,

53:05 other income filing,

53:06 disclosure documents,

53:08 and see if there are any suspicious transactions that are taking place,

53:11 any transactions taking place.

53:12 Is above a certain currency level,

53:14 that might indicate that there's something suspicious going

53:17 on in terms of their actual take-home income,

53:20 which might be out of line with

53:22 their

53:23 core job within the government sector

53:25 and the salary that they should be getting typically from that government job.

53:29 Um,

53:29 this is also something that you can do for your existing employees,

53:33 right?

53:33 So,

53:34 You can look at an employee,

53:36 um,

53:36 this person in your HR system is noted for this position title,

53:40 working in this grade in this ministry,

53:42 and this is the actual salary payment that was given out this month

53:47 to this person.

53:48 And if that is out of line,

53:50 if

53:50 the system is indicating that this person according to the grade

53:53 and title should actually be eligible for only this much amount

53:57 or.

53:57 They should be eligible for XYZ allowance,

54:00 but they're also getting

54:01 two other allowances which they're not technically eligible for.

54:04 The system could flag these and then you could go

54:07 in and actually make sure that that system is cleaned up

54:10 and that employee no more receives those

54:12 salary payments or those allowance payments.

54:14 Um,

54:15 and then finally some of the other things that we

54:17 found that could be very easily applied to HR systems.

54:21 Are things that,

54:22 um,

54:22 you know,

54:22 you could just have this system running these iterative tests

54:25 and you could see what is somebody,

54:27 um,

54:27 has somebody indicated that their home address is actually a PO box or a mailbox?

54:31 Are they receiving multiple paychecks?

54:33 Are they getting a paycheck or a bonus,

54:35 um,

54:35 that's out of line with sort of typically when this category of payment is paid.

54:39 Out

54:40 and the system could very well generate that.

54:42 These are the kinds of things that are very,

54:44 um,

54:44 would take a lot of time and effort if a human being had to do it,

54:48 but if you design the system to do it,

54:49 it can be done very cost effectively,

54:52 very easily,

54:53 and the system might be able to identify

54:55 many more such instances than a person could.

54:59 And this could also be applied to things like travel expense reports,

55:02 overtime checks,

55:02 etc.

55:04 uh,

55:04 really just any category of payments that are being paid out to workers,

55:08 um,

55:08 where the system could tag these very effectively.

55:11 So that's sort of,

55:12 um,

55:12 kind of examples,

55:14 there are a lot more that could be applied.

55:16 Um,

55:16 and again,

55:17 what you'll see here,

55:18 um,

55:18 if you go through the report,

55:19 we've provided country case studies where they are available,

55:23 um,

55:23 but a lot of this is hypothetical because it's not being applied to the HR

55:27 sector yet.

55:28 However,

55:29 given the state of digital technologies right now,

55:31 there's a lot more that can be done,

55:33 even in the HR sector,

55:35 and I think it's important to realize that.

55:37 You know,

55:38 when you're looking at big

55:39 bidding,

55:40 procurement,

55:40 if the systems,

55:41 those are important.

55:42 But these kinds of activities where somebody's not showing up to work,

55:46 collecting salaries and payments that they're not eligible for,

55:49 um,

55:50 when they're lying,

55:50 say,

55:51 or providing false records for recruitment and promotion,

55:54 these are all very important dementias as well that the government should

55:57 address and they affect the overall functioning of the public sector.

56:01 So,

56:02 um,

56:02 just the last bit that I want to spend maybe a couple

56:04 of minutes on because I'm mindful of the time as well.

56:07 Are some of the analog complements to digital reform,

56:10 um,

56:11 Hunt

56:11 kind of delved into these,

56:13 so I don't want to take up too much time,

56:15 but I wanted to just quickly flag some,

56:17 they're also very relevant for HR systems,

56:19 not just procurement systems,

56:21 right?

56:21 So,

56:21 data availability is a very big one,

56:25 We talked a lot about HR systems and cross-reference in data registries,

56:29 but if the data doesn't exist,

56:31 the government doesn't have an ability to do any of that.

56:34 Um,

56:34 and in fact,

56:34 what we find a lot of the public sectors that we actually work with,

56:38 um,

56:38 as part of our work with the World Bank is

56:40 that often public sectors do not have very comprehensive,

56:44 accurate,

56:44 timely,

56:45 centralized data systems.

56:46 So certainly that's something.

56:48 To

56:48 begin working towards and then also in parallel,

56:51 you can develop these digital systems

56:53 to detect and prevent fraud and corruption.

56:55 Data protection,

56:56 of course,

56:56 is also very important and Hunt alluded to this.

56:59 Um,

57:00 you know,

57:00 individuals have an intrinsic right to data privacy,

57:03 and therefore,

57:04 the government has to really exert an effort to make sure that they identify,

57:08 you know,

57:09 where,

57:09 uh,

57:10 an employee's.

57:10 Data will be used for what purposes,

57:13 by whom.

57:14 You have to make sure that there is effective cybersecurity,

57:17 um,

57:18 you know,

57:18 protocols to

57:19 prevent these threats,

57:20 prevent hacking instances and cyberattacks,

57:23 um,

57:24 make sure that you have the digital tools that you're applying,

57:27 but that government employees then don't misuse these digital tools

57:31 to go after their own opponents.

57:33 And then also to make sure that there aren't inherent

57:36 biases that are creeping into your algorithms and data protocols.

57:40 And this is where you have to constantly

57:42 retrain your algorithms and your digital tools.

57:45 Um,

57:45 and then of course,

57:46 there's data infrastructure and data sharing,

57:48 Han talked about this,

57:50 the public sectors that we work in,

57:52 um,

57:52 the culture is one of working in silos,

57:54 not really sharing information,

57:56 not really working together,

57:57 and so when you start talking,

57:58 About these data registries that need to be cross-referenced,

58:02 um,

58:02 across government,

58:03 across sort of outside of government as well.

58:06 What you really have to come up with are these data sharing protocols,

58:09 you have to assign roles,

58:10 make sure that business processes in place.

58:13 Um,

58:13 again,

58:14 who shares data with whom,

58:15 for what purpose,

58:17 in what manner,

58:17 and how that

58:18 will overall help them in their fight against,

58:21 uh,

58:21 corruption and fraud,

58:23 um,

58:23 data.

58:24 Literacy,

58:24 digital literacy,

58:25 and institutional capabilities again then come up.

58:28 Um,

58:28 do you have the right

58:30 skill set within your public sector for your staff

58:33 to be able to use these digital tools effectively?

58:36 And also,

58:36 are you creating a pipeline that you in the future,

58:39 have such digitally savvy employees in place?

58:43 And this would really require working with academic institutions and making sure

58:46 that the local curriculum is addressing that.

58:49 And then finally adapting to the local context,

58:51 um.

58:52 Again,

58:53 this is something Hunt talked about in relation to procurement and IMM systems,

58:57 um,

58:58 similarly relevant for HR systems as well.

59:00 Um,

59:01 what we found in some of the analysis that we did was that,

59:04 um,

59:05 you know,

59:05 where you already know that there is risk of corruption,

59:08 where you already know that this is typically where,

59:11 you know,

59:12 and this is what we've done,

59:12 so,

59:13 ghost workers,

59:14 somebody siphoning off like additional cash allowances

59:18 when they don't,

59:18 they aren't eligible for that,

59:20 typically.

59:21 In any government around the world,

59:23 people have an understanding of what the

59:26 risks are for fraud and corruption activities.

59:29 And so you should design a system that adapts to that and helps you track that,

59:33 um,

59:34 instead of just importing sort of this comprehensive digital system

59:37 that might not address the root causes of the

59:39 problem as they exist in your local context.

59:42 And finally,

59:43 linking to government decision making is intrinsic,

59:45 right?

59:45 So you could have this very fancy digital system,

59:48 you could have all kinds of.

59:49 Algorithms in place,

59:51 um,

59:51 you could be using it and you could be tracking it,

59:53 but if nobody's taking any action on these reports,

59:57 um,

59:57 if nobody's sanctioning employees as a result of their fraudulent activities,

1:00:02 nothing will get solved.

1:00:03 And so here you might have to develop certain regulations,

1:00:06 um,

1:00:06 and civil service procedures

1:00:08 to make sure that employees can in fact be sanctioned

1:00:11 for their fraudulent activities.

1:00:13 Um,

1:00:14 so that's sort of it for my presentation.

1:00:16 Thank you,

1:00:16 and let me stop here.

1:00:18 Thank you,

1:00:19 Isa,

1:00:20 and uh just to,

1:00:21 to thank both presenters uh for their excellent uh coverage of the issues

1:00:26 and maybe to also make an

1:00:29 observation that the report is rich in these case studies

1:00:33 and these examples that I think are really

1:00:35 uh helping to drive home some of these key messages.

1:00:39 Um,

1:00:39 one of the points.

1:00:40 I just made at the end reminded me of the example of Brazil

1:00:45 where the Comptroller General introduced AI to detect

1:00:49 who was most likely to be a corrupt civil servant,

1:00:52 but the law

1:00:54 hasn't adjusted to allow such things to be evidenced.

1:00:56 So

1:00:57 please do read the report because these examples

1:01:00 really illustrate some of the points being made.

1:01:02 Let's turn now to our discussions.

1:01:04 Uh,

1:01:05 first turn to Donna Andrews,

1:01:07 who's a senior public sector specialist,

1:01:09 and,

1:01:09 uh,

1:01:10 then we'll come to Helena Amaniova,

1:01:12 uh,

1:01:12 who's a senior operations analyst

1:01:14 for their key points before coming back to the Q&A.

1:01:17 Donna,

1:01:18 the floor is yours.

1:01:19 OK,

1:01:20 um,

1:01:20 thank you,

1:01:21 Tracy.

1:01:21 Um,

1:01:22 and first,

1:01:23 um,

1:01:23 thanks to Hunt and Izzo for some really interesting presentations.

1:01:27 There's,

1:01:28 um,

1:01:28 plenty of information and food for thought in there.

1:01:32 And I must say that this topic um I guess takes me back to a life before the World Bank,

1:01:38 um,

1:01:39 where I was actually the head of an ethical

1:01:41 standards unit in a public service commission in Australia.

1:01:44 And so I know that the issue of managing fraud from staff

1:01:48 is a universal challenge regardless of where you are in the world.

1:01:52 So I'm certainly glad that the bank is making a contribution to

1:01:55 and challenging the ways in which governments can tackle this issue.

1:02:00 Um,

1:02:00 I should probably also,

1:02:02 um,

1:02:02 I guess confess that I come from a HR background,

1:02:05 so I'm,

1:02:06 I guess somewhat biased about the importance of,

1:02:09 of HRM and the contribution that it can make to government effectiveness.

1:02:13 And so I'm going to focus,

1:02:14 um,

1:02:15 my comments largely on,

1:02:16 um,

1:02:17 ISA's HRM,

1:02:18 um,

1:02:18 presentation.

1:02:20 Now I think the,

1:02:21 the timing of looking more closely at how technology can help to identify HR fraud

1:02:26 really couldn't come at a better time.

1:02:29 So in most of our client countries,

1:02:31 um,

1:02:31 wage bill payments form a significant part of government.

1:02:35 Expenditure

1:02:36 and the ongoing impact of COVID is going to mean that the

1:02:39 government expenditure is likely to be under more pressure than ever.

1:02:43 So it's,

1:02:43 it's really critical that every dollar is spent well

1:02:47 and that it gets to its intended destination.

1:02:51 And the first thing I wanted to comment on and

1:02:53 perhaps this speaks more I guess to my HR background is

1:02:57 um and and Isa has talked about this,

1:02:59 is about the underutilization of HR data.

1:03:02 Um,

1:03:03 you know,

1:03:03 HR captures a huge volume of data about employees,

1:03:07 so everything from um date of birth,

1:03:10 family members,

1:03:11 national ID numbers,

1:03:13 remuneration,

1:03:14 qualifications,

1:03:15 banking information,

1:03:16 really the,

1:03:17 the list goes on,

1:03:18 it's a huge volume.

1:03:20 So there is um a requirement to collect it and to store it,

1:03:23 but generally then it's not referred to again

1:03:26 unless possibly um a problem arises somewhere else.

1:03:30 So in comparison to the procurement and financial information,

1:03:35 there really isn't a strong history of using technology and

1:03:38 data from HRM systems to help find fraud and misconduct.

1:03:42 In fact,

1:03:43 I I don't think I'm,

1:03:44 I'm too out of line to say that there really

1:03:47 isn't a strong history of using data from HRM systems,

1:03:50 even to inform decision making and planning more broadly.

1:03:54 So it's definitely an area,

1:03:55 um,

1:03:56 particularly where HRM professionals can and and need to do better.

1:04:01 The issue of um technology and data

1:04:04 analysis as key competencies for HRM professionals,

1:04:08 um you'll find is a key feature of um competency frameworks in

1:04:13 almost all advanced economies and most OECD countries.

1:04:16 So there's a recognition that we need to do

1:04:19 more and to help HR professionals to understand,

1:04:22 um some of the the key skills and experience that they need in that area.

1:04:28 But I think that um ISA provided some great examples,

1:04:31 but a stronger focus on using data from HR systems

1:04:35 can certainly help to avoid some very costly fraud situations,

1:04:39 um and particularly I think the example that that springs to my mind is,

1:04:43 um recruiting individuals without the necessary,

1:04:46 Qualifications.

1:04:47 Um,

1:04:47 I can recall a number of specific examples from my time in government in Australia,

1:04:53 where staff members were recruited to

1:04:55 positions which required specific qualifications,

1:04:58 only to discover at a later time

1:05:00 that the individual didn't have those qualifications.

1:05:04 Um,

1:05:04 and I,

1:05:05 I can recall an example um of a doctor

1:05:08 who was in fact recruited to a regional hospital.

1:05:11 And the issue about the qualifications only arose after a

1:05:15 series of complaints about poor patient care and outcomes.

1:05:19 And I think in times of COVID,

1:05:20 ensuring that appropriately qualified staff are

1:05:23 being engaged for vaccine programs,

1:05:25 for example,

1:05:26 it seems like a really critical decision

1:05:28 and one where using technology to validate qualifications

1:05:32 could certainly save significant time and money.

1:05:36 And the US Department of Labor recently estimated that the cost of poor

1:05:40 recruitment decisions can be up to 30% of an individual's first year salary.

1:05:46 Now that can be a significant amount of money,

1:05:48 um,

1:05:49 and I think that that that money could certainly be better invested in perhaps

1:05:53 learning and development for staff or even um improving IT systems.

1:05:58 So I wanted to um just add a couple of further examples in terms of

1:06:02 recruitment and selection um on top of what ISA has has talked about already,

1:06:07 where technology can provide some great

1:06:10 opportunities for for fraud identification.

1:06:13 So I guess while technology has brought some great opportunities to streamline,

1:06:18 um,

1:06:18 sooner we can now use AI to screen um CVs and job applications,

1:06:24 we can use one-way interviews where candidates can respond

1:06:27 to pre-recorded interview questions during a timed online interview.

1:06:32 Uh,

1:06:32 we can use chatbots to um

1:06:34 conduct initial screening questionnaires with candidates.

1:06:37 But it's also brought the opportunity for some candidates to use technology

1:06:41 to try to gain an advantage or seek to deceive recruiters.

1:06:45 Um,

1:06:46 so particularly in the private sector,

1:06:48 organizations are adopting some,

1:06:50 some high-tech solutions to help better manage the risks.

1:06:54 So for example,

1:06:55 um,

1:06:56 looking at using blockchain technology to identify employees,

1:07:00 um,

1:07:01 individuals who have,

1:07:02 um,

1:07:03 provided fake qualifications,

1:07:05 uh,

1:07:05 using facial recognition in video interviews to ensure that the

1:07:09 applicant is actually the one who is being interviewed.

1:07:12 Assessing voice and speech tone

1:07:15 and using some sophisticated written assessment tools to

1:07:18 ensure the the authenticity of candidates claimed skills,

1:07:22 so there is a huge volume of these

1:07:24 and of course they're evolving and moving forward um

1:07:28 um continually.

1:07:30 The third issue I wanted to touch on just briefly is

1:07:33 the analog complements that that Isa Anne Hunt referred to.

1:07:37 Um,

1:07:38 for me,

1:07:38 I think one of the key takeaways is really the the

1:07:41 confirmation that the digital tools aren't a be all and end all

1:07:46 just for themselves.

1:07:47 We really need to combine the digital reforms with the analog complements

1:07:52 in order to be able to make the most out of the technology.

1:07:56 Um,

1:07:56 and I think one issue which is worth underscoring,

1:07:59 um,

1:07:59 again that that is a raised is about the staff skills for technology.

1:08:03 So the effective use,

1:08:05 being able to effectively use HR data

1:08:08 requires staff who are familiar with the HR systems,

1:08:11 but also the type of data that's available,

1:08:14 but also that they've got the data analysis

1:08:17 skills in order to make sense of the information

1:08:20 and to be able to use that as part of identifying potential problems or issues.

1:08:26 So traditional fraud and misconduct investigations um

1:08:29 definitely still rely on the human dimension,

1:08:33 but technology can play a great part in helping to

1:08:36 point us to areas where we should look deeper,

1:08:39 well before we,

1:08:41 we might otherwise know that a problem exists.

1:08:44 Now we know within,

1:08:46 within the World Bank from surveys of civil

1:08:48 servants that many governments face some basic constraints,

1:08:52 including a lack of digital skills.

1:08:54 Um,

1:08:54 for example,

1:08:55 in the Philippines,

1:08:56 which is a middle-income country with a fairly vibrant digital economy,

1:09:01 um,

1:09:01 only 8 out of 10 staff could in fact create a PowerPoint presentation.

1:09:06 And if we look at um surveys from Ethiopia,

1:09:10 in fact,

1:09:10 fewer than 50% of staff in government could use a computer

1:09:14 to write a memo or to create an Excel spreadsheet.

1:09:17 So given the speed at which technology is moving,

1:09:20 there certainly needs to be continual focus on staff skill development.

1:09:25 And in fact within the bank we're currently working on a report which will focus on

1:09:29 gov tech skills in the civil service and how to build and and retain those skills.

1:09:35 So finally,

1:09:36 um,

1:09:36 let me um thank the team for a great and really practical report

1:09:40 and some excellent examples of,

1:09:42 of how technology can help,

1:09:45 um,

1:09:45 with fraud detection in public

1:09:47 administration.

1:09:48 I certainly

1:09:49 encourage you all to,

1:09:50 to download and to read that.

1:09:52 And perhaps let me just um finish um with um a quick question for ISA,

1:09:57 um,

1:09:58 which hopefully we might have some time to to look at is,

1:10:01 given that particularly in HRM the use of data and technology is

1:10:05 really in its infancy in any of the countries that we work,

1:10:09 um,

1:10:09 what are your suggestions about how countries could

1:10:11 get started on this type of agenda,

1:10:14 particularly in um the HR area?

1:10:16 So thanks very much.

1:10:19 Thanks,

1:10:19 Donna.

1:10:19 Let's move quickly to Irina who's been very patient,

1:10:22 uh,

1:10:22 waiting for her turn.

1:10:24 Irina,

1:10:24 over to you.

1:10:34 You're the

1:10:35 sound

1:10:35 quality is not working.

1:10:42 Elena,

1:10:42 can you um try to

1:10:44 mute and then open or?

1:10:49 Microphone again

1:10:57 Oh,

1:10:57 we,

1:10:58 we cannot hear you.

1:10:59 Um,

1:11:02 Actually,

1:11:03 I suggest that you can you can you um

1:11:06 turn off the the um the blur,

1:11:08 the blur effect.

1:11:10 It might help

1:11:12 So why don't we let Irina and Natalie solve

1:11:15 the audio problem and whilst we're doing that,

1:11:19 um,

1:11:19 I will pick up a couple of the questions,

1:11:22 um,

1:11:22 back to the presenters.

1:11:24 Um,

1:11:25 Natalie,

1:11:26 maybe you could read Irina for a minute.

1:11:30 And so the first question back to the team,

1:11:34 uh,

1:11:36 what,

1:11:36 what's the actual evidence on the financial management system's

1:11:41 improvements when it comes to issues of fraud and corruption?

1:11:46 A question from our colleague in Malawi says

1:11:49 that

1:11:49 there's a lot of perception,

1:11:51 but in fact,

1:11:53 Incidences of fraud in payments seems to have increased

1:11:56 with the implementation of online financial management systems.

1:12:00 Maybe we could turn just to that question,

1:12:02 Hunt,

1:12:03 and there's another question in the chat for you

1:12:07 if Irina doesn't come back online,

1:12:09 which is

1:12:10 how is potential fraud monitored

1:12:12 during project implementation.

1:12:14 Over to you,

1:12:14 Hunts.

1:12:16 Great,

1:12:16 thank you so much.

1:12:17 Now,

1:12:18 um,

1:12:18 one of the,

1:12:19 uh,

1:12:20 regarding,

1:12:21 uh,

1:12:21 fraud and payment systems and uh the detection

1:12:25 of digital detection of fraud and payment systems,

1:12:28 um,

1:12:29 in the

1:12:30 Uh,

1:12:30 in the countries where we work,

1:12:32 uh,

1:12:33 um,

1:12:34 in,

1:12:35 in IMIS systems,

1:12:36 very rarely are these fraud detection

1:12:39 systems,

1:12:41 uh,

1:12:42 installed and used.

1:12:44 Many times there is a political economy issue

1:12:48 with,

1:12:48 uh,

1:12:49 having,

1:12:49 um,

1:12:50 Uh,

1:12:51 getting them installed.

1:12:52 Uh,

1:12:52 so,

1:12:53 and additionally,

1:12:55 uh,

1:12:55 a capacity issue with getting them up and,

1:12:58 uh,

1:12:59 running and,

1:12:59 and monitoring,

1:13:00 monitoring them.

1:13:01 So,

1:13:02 I'm not aware at the moment of any,

1:13:05 uh,

1:13:05 implementation of these modules in,

1:13:08 uh,

1:13:08 ITMIS systems.

1:13:10 Um,

1:13:10 I'm more of a procurement expert.

1:13:12 Uh,

1:13:13 uh,

1:13:13 this is something we could ask,

1:13:15 uh,

1:13:15 some of our IMIS experts,

1:13:17 uh,

1:13:17 at the bank.

1:13:19 Um,

1:13:19 but they,

1:13:20 they are available,

1:13:21 uh,

1:13:22 and they are certainly used in OECD countries and,

1:13:26 uh,

1:13:26 in the private,

1:13:27 uh,

1:13:27 sector also.

1:13:29 Um,

1:13:30 so thank you for that question,

1:13:32 and I,

1:13:32 I think there,

1:13:33 there is a huge,

1:13:34 uh,

1:13:35 uh.

1:13:36 Uh,

1:13:36 a huge,

1:13:37 um,

1:13:38 possibility or opportunity,

1:13:40 if you will,

1:13:40 to uh incorporate some of these,

1:13:43 especially in ITMS,

1:13:44 uh,

1:13:44 uh,

1:13:44 some of these,

1:13:45 um,

1:13:46 uh,

1:13:46 fraud detection,

1:13:47 uh,

1:13:48 modules.

1:13:49 Um,

1:13:50 in,

1:13:50 in e-procurement,

1:13:51 for example,

1:13:52 what happens is,

1:13:53 is that,

1:13:53 you know,

1:13:55 Countries are,

1:13:56 are so focused on just digitizing the paper-based e-procurement system that,

1:14:01 you know,

1:14:01 they're,

1:14:01 they're not looking

1:14:03 for a fraud detection system.

1:14:04 They're looking for an e-procurement system which

1:14:07 helps them digitize the procurement process.

1:14:09 So,

1:14:10 um,

1:14:10 we're,

1:14:11 you know,

1:14:11 we're thinking of,

1:14:12 of,

1:14:13 of enhancing uh the

1:14:16 uh the,

1:14:16 the e-procurement.

1:14:17 Um,

1:14:18 I guess,

1:14:18 uh,

1:14:19 projects that,

1:14:20 you know,

1:14:20 phase phasing it out,

1:14:22 uh,

1:14:22 and adding an integrity module,

1:14:26 uh,

1:14:26 and later phases of a new procurement system,

1:14:29 um,

1:14:29 regarding the,

1:14:30 uh,

1:14:31 how do we,

1:14:32 how do we monitor,

1:14:33 uh,

1:14:34 fraud detection,

1:14:35 uh,

1:14:36 in,

1:14:37 in,

1:14:37 in,

1:14:37 uh,

1:14:37 bank projects?

1:14:38 Well,

1:14:38 it's interesting,

1:14:39 um.

1:14:40 Uh,

1:14:42 it was very,

1:14:43 it was,

1:14:43 uh,

1:14:45 before

1:14:45 the new procurement framework,

1:14:47 uh,

1:14:47 was adopted in,

1:14:49 uh,

1:14:49 July 1st,

1:14:50 2016.

1:14:51 It was per transaction,

1:14:54 right?

1:14:54 If,

1:14:54 if a procurement specialist who was doing a prior review saw

1:14:58 something,

1:14:59 some hanky panky,

1:15:00 uh,

1:15:01 in the bids during the evaluation process and the evaluation report,

1:15:05 then,

1:15:05 uh,

1:15:06 we would,

1:15:07 we would take a look at it and if it looked as if,

1:15:10 uh,

1:15:11 there could be some fraud there,

1:15:12 we would send it to INT and INT would do the due diligence of following up on that.

1:15:18 So that was historically how it was done and that's how it's done

1:15:22 now on an individual basis,

1:15:24 um.

1:15:25 Now that we have SEP,

1:15:26 we have a lot of useful information in SEP,

1:15:29 and I know that our colleagues in INT and ISP,

1:15:33 uh,

1:15:33 are,

1:15:34 are building a monitoring,

1:15:36 uh,

1:15:37 uh,

1:15:38 algorithms that can look at this data.

1:15:41 Now,

1:15:41 SEP is really just a,

1:15:42 a,

1:15:43 a procurement monitoring system.

1:15:45 It is not,

1:15:46 uh,

1:15:46 an e-procurement system,

1:15:48 but there's a lot of very useful information in there to help,

1:15:51 um.

1:15:52 Uh,

1:15:53 help,

1:15:53 uh,

1:15:53 do that.

1:15:54 So that would be my,

1:15:55 at the moment,

1:15:56 we do it on an individual transactional basis and still somewhat manual.

1:16:00 Hopefully in the near future it will become,

1:16:02 uh,

1:16:03 much more aggregate,

1:16:04 uh,

1:16:04 with other fraud detections.

1:16:05 Thank you.

1:16:06 Thanks,

1:16:07 Hunt.

1:16:07 Irina,

1:16:08 let's,

1:16:08 uh,

1:16:08 let's try again to,

1:16:10 to connect you.

1:16:20 We,

1:16:20 we,

1:16:21 we still cannot hear you.

1:16:24 Can you try turning off the video?

1:16:26 That might help.

1:16:32 Can you try um your audio?

1:16:36 Um,

1:16:37 OK.

1:16:40 OK,

1:16:40 Irina,

1:16:41 I'm afraid we still don't hear you well,

1:16:45 um,

1:16:46 unfortunately,

1:16:47 I know that this did work in the,

1:16:50 the,

1:16:50 the

1:16:51 beginning of the session,

1:16:52 so something happened over time.

1:16:54 Maybe Irina,

1:16:55 you could,

1:16:56 um,

1:16:56 reconnect through another device,

1:16:59 um,

1:16:59 and maybe

1:17:01 leave this one on for video and use a different device for audio.

1:17:09 So,

1:17:09 uh,

1:17:10 turning to the next question in the chat,

1:17:12 I'm going to paraphrase it a little bit,

1:17:14 um,

1:17:14 but as we use

1:17:16 more sophisticated tools to detect

1:17:20 wrongdoing or bad behavior,

1:17:23 how long is it before

1:17:25 the people who would like to circumvent these,

1:17:29 what is it,

1:17:29 they're circumventing

1:17:30 manual processes now,

1:17:32 get smart enough to circumvent the e-tools or the new

1:17:36 IT systems that you're bringing in.

1:17:38 So how long does it

1:17:39 Take before somebody knows how to go into the system and,

1:17:42 and just talk to the,

1:17:43 you know,

1:17:43 the qualifications,

1:17:45 um,

1:17:45 or to,

1:17:46 you know,

1:17:47 ensure that the red flag

1:17:48 isn't shown up in the procurement system.

1:17:50 So question back to you,

1:17:52 is it a hunt

1:17:53 as to whether or not this,

1:17:54 uh,

1:17:55 whether or not the government can always stay one step ahead or not.

1:18:00 I said,

1:18:00 do you want to go first?

1:18:02 Sure,

1:18:02 let me go first and then,

1:18:03 um,

1:18:04 Hunt,

1:18:04 please feel free to add on to what I say.

1:18:07 I mean,

1:18:07 this is a very good point,

1:18:09 right?

1:18:09 And,

1:18:10 um,

1:18:10 I think the fight against perhaps fraudulent behavior will be.

1:18:15 Always there,

1:18:16 uh,

1:18:16 as much as we might not want to see it,

1:18:19 uh,

1:18:19 but I think that's why the report focuses not only on

1:18:23 describing what these digital tools can look like and what the system can look like,

1:18:26 but also emphasizes the analog complements.

1:18:29 Um,

1:18:29 and this is really where you can try to make sure that as people then

1:18:34 try to game the digital tools themselves,

1:18:37 the government has systems in place to make sure,

1:18:40 um,

1:18:40 that that does not happen.

1:18:41 So things,

1:18:42 for example,

1:18:43 like,

1:18:44 You know,

1:18:45 uh,

1:18:45 talk,

1:18:45 we talked about digital literacy,

1:18:46 but also promoting a culture that is ethical.

1:18:49 I mean,

1:18:50 it's not just about digital systems that help

1:18:52 you catch somebody after they've done something.

1:18:55 It's also to actually create and promote a very ethical culture

1:18:59 within the public sector where you have guidelines that say,

1:19:02 look,

1:19:02 like this is what we stand for,

1:19:03 these are our values,

1:19:04 this is what we don't stand for,

1:19:06 and this is why

1:19:08 behavior that is unethical is not only bad for business.

1:19:11 But it's also something that we don't believe in as a,

1:19:13 as public servants.

1:19:15 So,

1:19:15 I think that culture also plays a very important role on its own,

1:19:19 and government should put in place,

1:19:21 um,

1:19:21 you know,

1:19:21 training modules,

1:19:22 guidelines,

1:19:23 value systems,

1:19:24 trainings that can help promote that.

1:19:27 But in parallel,

1:19:28 of course,

1:19:28 you should also develop systems that help you stem

1:19:31 these kinds of unethical behaviors as they take place.

1:19:34 Um,

1:19:34 and this is where I would like to refer to something that Donna mentioned,

1:19:37 which is a very good point.

1:19:39 Which is that,

1:19:40 again,

1:19:40 digital tools are not a panacea.

1:19:42 You have to pair it with human intervention.

1:19:44 So,

1:19:45 um,

1:19:45 it's about having still the auditors in place,

1:19:48 having still the investigators in place,

1:19:49 having still the regulators in place,

1:19:52 um,

1:19:52 making sure that they use technology as a tool that helps them

1:19:57 address fraud and corruption much more effectively,

1:19:59 but still,

1:20:00 Having that human in place who can try to stem the tide and kind of be one step ahead

1:20:05 of people who might still try to misuse the system.

1:20:07 So,

1:20:08 I think for me,

1:20:09 the idea is use technology as a tool,

1:20:11 still have the human intervention,

1:20:13 and then make sure that you're also working towards

1:20:15 expanding the ethical culture within the public sector.

1:20:18 But,

1:20:18 um,

1:20:19 Hunt,

1:20:19 any additional points that you would like to add?

1:20:22 Uh,

1:20:23 I,

1:20:23 I would,

1:20:23 I would say,

1:20:24 you know,

1:20:25 uh,

1:20:25 it's a great point because in the future there's gonna be different

1:20:28 types of fraud,

1:20:29 but let's focus on the,

1:20:31 the,

1:20:31 the ones we've identified in this report,

1:20:33 we have,

1:20:34 uh,

1:20:35 ISA has,

1:20:35 has presented,

1:20:36 uh,

1:20:37 an,

1:20:37 an amazing set of,

1:20:39 um,

1:20:39 schemes,

1:20:40 uh,

1:20:40 and algorithms.

1:20:41 Um,

1:20:42 that can identify,

1:20:44 uh,

1:20:44 fraud that's happening right now.

1:20:46 We also have this for IMS,

1:20:47 and we have it very well spelled out for procurement.

1:20:50 So focusing on those now can really

1:20:53 enhance,

1:20:53 and it's,

1:20:54 and it's using the digital side of this,

1:20:57 uh,

1:20:57 with an aggregate,

1:20:58 uh,

1:20:59 look at what's happening.

1:21:01 Um,

1:21:01 and,

1:21:02 and I think that the speed of once it becomes all digitized,

1:21:06 uh,

1:21:07 it,

1:21:07 it,

1:21:07 it,

1:21:08 it,

1:21:08 it's gonna happen pretty quick.

1:21:10 Um,

1:21:10 there will be new fraud schemes,

1:21:11 they always come up,

1:21:12 um,

1:21:13 but I think the speed of attacking those is

1:21:15 gonna happen much faster once everything becomes digitized.

1:21:19 Thanks.

1:21:20 Thanks,

1:21:20 Susa,

1:21:20 thanks,

1:21:21 Hunt.

1:21:21 Irina,

1:21:22 let's,

1:21:22 uh,

1:21:22 let's turn to you.

1:21:24 Ah,

1:21:25 thank you.

1:21:26 So,

1:21:28 now you can hear me,

1:21:29 I hope.

1:21:31 Thank you very much.

1:21:32 Uh,

1:21:33 so first of all,

1:21:34 uh,

1:21:34 thanks a lot for this great presentation and

1:21:38 for the information that you provided to us.

1:21:41 Uh,

1:21:41 I would like to focus on fraud

1:21:43 in procurement because I have a procurement background

1:21:47 and now I work in OPCS anti-corruption program.

1:21:52 Which is also analyzing fraud and corruption cases in our operations,

1:21:58 and

1:22:00 it is

1:22:02 very,

1:22:02 really important to

1:22:04 identify fraud.

1:22:07 At the earliest stage,

1:22:11 so the tools which propose ex ante

1:22:16 ex ante measures to identify fraud

1:22:19 is very important indeed.

1:22:22 For instance,

1:22:23 our analysis of fraud and corruption cases in operations shows that

1:22:29 fraud cases

1:22:31 are just Fraud at the bidding stage

1:22:34 is a prevailing type of fraud and corruption practices,

1:22:40 and it occurs either on its own or

1:22:43 in combination with other fraud and corruption practices.

1:22:47 And

1:22:50 ours,

1:22:50 if

1:22:51 they are trained

1:22:52 to detect this fraud at the billing stage,

1:22:56 they We can effectively identify it and

1:22:59 prevent fraud and corruption in our projects and

1:23:05 in our contracts,

1:23:07 but unfortunately,

1:23:08 according to our information,

1:23:10 the can identify fraud

1:23:13 in our operations only about 50% of cases.

1:23:21 Training and

1:23:26 also sharing the information

1:23:29 on

1:23:31 on digital tools which can prevent fraud and corruption

1:23:36 is very important and

1:23:39 maybe we can also think how we can use this information which is presented today

1:23:45 and

1:23:46 which is described in.

1:23:48 The report.

1:23:49 How can we use this information

1:23:52 to share,

1:23:54 to share the tools,

1:23:55 the options which exist,

1:23:57 but also

1:23:59 to provide more information on governance filters that you can apply

1:24:04 to detect fraud and corruption?

1:24:07 How can we share this with our operations

1:24:11 with procurement specialists

1:24:13 so that they put

1:24:15 Place some mitigation measures

1:24:18 and how can we also use

1:24:22 these

1:24:24 automated systems in our operations if it is

1:24:28 possible

1:24:31 to use at least some modules

1:24:33 in our operations to help the

1:24:37 to prevent fraud and corruption cases in our operations

1:24:42 and

1:24:44 Also,

1:24:45 uh,

1:24:46 uh,

1:24:48 I think that

1:24:50 it is very important to understand that

1:24:54 some of the modules can be used ex ante and to prevent front and corruption,

1:24:59 but

1:25:00 other modules can be used

1:25:03 post factum

1:25:05 because you need some more information to collect information from.

1:25:09 Procurement systems to see some,

1:25:13 for instance,

1:25:14 collusive patterns

1:25:16 or to see

1:25:18 some

1:25:19 fraud at the contract implementation stage

1:25:24 so that some of the modules can be used by procuring entities,

1:25:28 but other modules.

1:25:30 Can be used by

1:25:32 internal or external

1:25:35 inspections and audits.

1:25:37 So

1:25:39 I'm sorry.

1:25:40 I understand I don't have much time because

1:25:44 we are approaching the end of our session,

1:25:47 and I'm sorry that I had these issues with connection.

1:25:51 But once again,

1:25:53 I would like to thank the presenters and

1:25:56 I would encourage everybody to read the report because

1:26:00 you have much more information which is very important and

1:26:04 which can be used in the policy dialogue,

1:26:08 but also

1:26:09 some of this information we can use in our

1:26:13 operations to prevent fraud and corruption in our projects.

1:26:16 Thank you very much.

1:26:18 Thank you,

1:26:19 Irina,

1:26:19 and thank you for persevering with the technology this morning.

1:26:22 It was great to have your,

1:26:23 your comments at the end here.

1:26:26 Uh,

1:26:26 let me close the session.

1:26:28 I think there's a couple of requests for additional references to reports,

1:26:32 um,

1:26:32 that,

1:26:33 that

1:26:33 the presenters and discussions made.

1:26:35 So we'll make sure they go out with Natalie and the recording of the session.

1:26:38 Uh,

1:26:38 let me thank everybody,

1:26:40 uh,

1:26:40 for attending.

1:26:41 This has been one of the most well attended,

1:26:44 uh,

1:26:44 governance.

1:26:45 Uh,

1:26:45 BBBS this year.

1:26:47 So thank you to everyone.

1:26:49 I think that's a good testimony to the

1:26:51 team and their topic and their presentation that these

1:26:55 100 participants are still here with us at the end of a couple of hours.

1:26:59 So

1:27:00 thank you to you all.

1:27:01 Thank you to those who've all been working on the report.

1:27:04 And let me just take this moment to flag some upcoming relevant

1:27:08 pieces of work because as Donna mentioned,

1:27:10 we have

1:27:11 new guidance.

1:27:12 Coming on how to develop gov tech skills in the public sector

1:27:16 and on the procurement side,

1:27:18 the team have been working on a

1:27:19 prototype for detecting uh fraud using artificial intelligence

1:27:24 in procurement systems.

1:27:25 So we'll be hearing more on both these topics in the future.

1:27:28 Uh,

1:27:28 thanks again

1:27:29 to,

1:27:30 to the team.

1:27:31 Thanks to Ed for joining us this morning,

1:27:33 and I'm wishing everybody

1:27:34 a good evening,

1:27:35 good rest of the day,

1:27:37 and,

1:27:37 uh,

1:27:37 see you at the next one.

1:27:38 Thank you.

1:27:39 Thank you all.

1:27:40 Goodbye.

1:27:42 Thank you.

1:27:43 All the best.

1:27:44 Thank you so much.

showAllTimestamps

no

transcript

Great. Good morning, good afternoon and good evening, everyone. It's great to see so many of you connected. We're at 78 attendees this morning, uh, afternoon. Great that you were able to, to join us. This is the latest, uh, BBB Round Bag brunch or breakfast or dinner wherever you are, uh, from the GovTech series, and we're trying to vary our time so that we can have more attendance from the field. Our topic today is the use of technology to find fraud and corruption in the public administration. The United Nations estimates that $1 trillion is paid in bribes, $2.6 trillion lost to corruption every year, and combined this is 5% of global GDP. In low income countries, the cost of corruption is estimated to be about 10 times the inward flows from ODA. In the current environment, COVID-19 has caused a greater perception of corruption being a problem, and we've seen lower trust in government institutions. So this new report is even more relevant today. Finding fraud, govtech, and fraud detection in the public administration examines new approaches to using technology, digital analysis of data, and artificial intelligence to detect and prevent fraud and corruption in the public administration. The focus of the report is on government systems for procurement, financial management, and human resource management. The report also addresses other legal, policy and political requirements for these tools to be successful, and the report's intended to be a practical guide for practitioners, policymakers and government officials, and for our TTLs to use in their policy dialogue. This morning we will have opening remarks from the governance GP Global director Ed Ololo Okeri, and then we'll hear from two of the experts on the report, Hunt Lacasia, a senior procurement specialist, and Iza Malik, a public sector specialist in the Menda region. And then we'll turn to our discussants this morning. So a warm welcome to Irina Smeliova and Donna Andrews, our discussants, and we will end with some Q&A from the panel. Please use the Q&A chat box and post your question there, and we will come to you at the end so you can post questions as we go through. Ed, over to you for welcoming remarks. OK, thanks, uh, Tracy. Uh, good morning, good afternoon and good evening to everyone. Uh, I'm delighted to join you in this event that is focused on the important issue of fraud. This is an age-long issue that continues to be with us, even in a period of crisis as we are currently. Over the years, various approaches and tools have been developed to prevent and detect fraud. This mostly build on rule of law and transparency as well as the classical internal and external control principles. In recent years, breakthroughs in digital technologies have expanded the horizon of possibilities and provided an array of new tools to governments to help them tackle fraud challenges and also address the broad issue of controlling corruption. GovTech, including e-government systems and e-services, present an exciting new frontier in the efforts to tackle the challenges of fraud and corruption. These tools That is e-government systems and e-services tools are closely related to improved outcomes in government effectiveness and perceptions of corruption. Some preliminary analysis by the Govtech team show that governments that rank high on UN Online Services Index. Perform better in corruption ranking and government effectiveness indicators, and this will not be surprising. It is really intuitive that if governments are using a lot of these services or digital methods to provide services to citizens, that will reduce interactions between human beings and also reduce the opportunities to be able to ask for and receive bribes. Also, if government is using digital tools to provide services, I think governments will be able to reach more people and we will be able to reach them in a more transparent ways. So we really think that these tools can be able to help in many ways to control fraud and corruption. As highlighted in the report that Tracy alluded to, which is being discussed this morning. There are significant opportunities for and benefits from using digital tools to tackle fraud and corruption challenges in the public sector. Then these are benefits that extend beyond just detecting and preventing fraud and corruption. In fact, studies show that taking advantage of the full potential of government digitization can free up to $1 trillion annually in global economic value through lowered costs and improved operational performance. At the same time, We also need to know that digital tools are not a panacea to all the challenges of fraud and corruption, as the country case studies in the report that Tracy mentioned illustrate. Technology is most effective when it is paired with traditional fraud detection and prevention methods, and they are integrated with analog components of the reform. I also want to emphasize that the rule of law is very fundamental to tackling corruption and fraud in any country context. I look forward to a very rich discussion today and I want to thank the organizers and also the speakers and the discussions. Thank you so much. Back to you, Tracy. Thanks, Ed. We're um moving straight to Hunt and, uh, Hunt, you have the floor of the, uh, the presentation as well. Thanks. Thank you very much, uh, Tracy and Ed, for the marvelous, uh, introduction. I'm going to, uh, share my screen at the moment. Let's see, and I will. There we go. OK. Hopefully, everybody can see uh the presentation. So, today, uh, the event on finding fraud, and this is based, uh, as Tracy and Ed, uh, discussed, uh, on a report called Finding Fraud GovTech and fraud detection in Public Administration. Um, And this really focuses when we talk about public administration we're talking about in this report 3 separate information systems in the public sector that's Uh, ImusS and HR, I will be discussing, uh, procurement and ITMIS, and our colleague Izza will be discussing, uh, the HR, uh, systems. Um, so I am, uh, a senior procurement specialist in the governance procurement global unit. Uh, I had the pleasure of TTLing this particular report, uh, and I look forward to going through, uh, and discussing. Uh, some of the great findings of this report. Um, we'll talk about, uh, you know, what the major issues are which Tracy and Ed alluded to, uh, what challenges exist, uh, and what, uh, digital tools can, uh, can countries deploy to combat corrupt practices and other drivers that should be considered, uh, what are some applications and benefits, and then key recommendations and next steps. Um, so, uh, thanks, Tracy and Ed for, uh, in, um, explaining why this is a major issue. One thing I did want to emphasize is that fraud and corruption worsens during emergencies, uh, and the increased demand and time pressures for acquisition of, uh, remedial goods and services, especially during COVID-19 has been a major, uh, issue. So, and they can lead to the relaxation of procurement and inspection. Creating an increased risk of selection of unqualified or fictitious suppliers and delivery of poor quality or non-existence, um, goods and services. So, uh, uh, just wanted to emphasize that in the introduction, um. And I will move right on to why, uh, what challenges exist. Uh, and, uh, so here we have, uh, some of the major challenges that do exist and that we do, uh, we do deal with and not, uh, no technology or digital tools. Can guarantee success when it comes to government reform efforts, right? Um, and studies have shown that there, uh, that there is a skills and resource gap when it comes to technology and, uh, digital. In the public sector, and political will, uh, it needs to be installed, uh, digit, uh, when you're, there needs to be political will, uh, when, uh, we have digital, uh, anti-fraud systems, what's one of the issues that we deal with, uh, and also there needs to be a follow-up on the results uh of on with appropriate, um, sanctions in the absence of many of the countries, uh. Uh, uh, where the systems are needed most. Also, poor infrastructure, including, uh, such as intermittent internet access. Uh, and lack of computing, uh, power, uh, to process the data and applications involved in procurement and IT systems. Also, poor financial management strategy can result in reduced usage and minimal standards and corrupt financial, uh, reporting practices, uh, and automation of bad practices and out of date institutions and even lack of translated materials in many of our, uh, client countries. Um, so these are, uh, uh, additionally, governments, uh, can work in developing countries can work in silos, and that's one of the issues that we deal with, uh, many times in the topic we're discussing today. Many times the public procurement agency, uh, Doesn't, uh, um, interface well with the, um, uh, with the Ministry of Finance, meaning that the information system, the e-procurement system for the public procurement agency isn't integrated with the data from the IMI system and therefore it can be very challenging, uh, to share information back and forth. Um, so what are the challenges in procurement? Uh, combating fraud and corruption in procurement, uh, is a central in digital, uh, detection. Procurement is where most fraud, uh, and corruption cases and losses occur, uh, where governments and where governments spend the most money, often, uh, financed by international donors. Um, but few e-procurement systems currently include ex-ante fraud detection programs, which we refer to as governance or integrity filters in the routine purchasing of uh, uh, in routine purchasing. And It appears that there's no such programs that monitor large scale Tinder transactions, which are serious losses where serious losses are routinely incurred. This could be in infrastructure projects and in large IT projects. And IMIS, uh, platforms are cited as useful in fraud detection. In fact, we'll, uh, demonstrate some very nice, um, or not demonstrate, but we'll, uh, discuss some, uh, IMIS modules that can be, uh, implemented. Uh, but IMIS in general is expensive, complex, and difficult to install and operate. Um, additionally, IMIS projects can have, uh, other issues including unsound project design and lack of necessary underlying finance reforms. Um, and I, uh, systems can be vulnerable to, uh, several fraud and corruption, uh, schemes such as the misallocation of budget items. Processing of inflated payments to shell companies or phantom vendors and payments to offshore accounts as part of money laundering schemes. Um, now we'll start talking about some of the digital tools and some of the, uh, fraud schemes and algorithms. Um, And in, in recent years, there has been a Uh, in recent years, there have, uh, been breakthroughs in digital technologies that have expanded the scope of reform, uh, possibilities and provided an array of new, uh, tools to governments to help them improve, uh, governance outcomes and control corruption. Um, and so, effectively use of digital tools, uh, requires service. Let me make sure I'm on the right page here. Yup, it requires uh services of high qualified, uh, uh, diligent and ethical development professionals, a proper training in anti-fraud measures and, and equipped with tools to detect and prevent it, such as, uh, the digital fraud tools we'll be discussing. And in GovTech, uh, which we have many e-government tools and e-services is closely related to improving outcomes and government effectiveness and perception of corruption. So, um, today we're gonna be talking about, uh, we're gonna start talking about some of the detailed integrity filters, which are sophisticated algorithms to detect and prevent fraud and irregularities that can be embedded in any procurement and IMI system. And these filters can run proactively, uh, which means, uh, ex ante, uh, to identify possible fraud over before, uh, bids are evaluated or payments are approved or ex post against procurement data stored in historical databases. Um, Uh, and here for our integrity filters, uh, we exploit, uh, many of the, uh, electronic procurement, uh, data that we have. So, here we can block non-compliant transactions. Uh, we can provide instant proactive alerts of possible fraud, uh, and more so with electronic procurement, we can review 100% of all. Transactions were previously uh during the paper-based processes you could only do one transaction at a time and it allows uh for a real-time remote monitoring by donors or oversight agencies if need be and there can be uh uh create a detailed audit trails and digital. Evidence for investigators, um, uh, and we also identify evidence of previous ongoing misconduct in, uh, in historic databases, and we have different types of Uh, levels of fraud, right? Uh, and, uh, in reports we have significant procurement statistics, uh, where these would be numbers of awards to certain contractors by certain approving officials. We also have reports on economy and efficiency indicators that ensure the selection of the best product at the best price, um, uh, compliance reports, obviously, uh, where contractors in violation of procurement rules. Can be identified, uh, maybe a short bid notice could be also identified, uh, or, or bids from a debarred, uh, company. We also have SPQQD reports, and these are select, uh, selection price, uh, qua quantity, quality, and delivery indicators that can point to fraud, uh, waste, or abuse. Uh, and then we also have, uh, information on collusive bidding and big rigging which we're gonna dive into. Um, and here are some of the common fraud and corruption schemes, uh, in procurement and that do reach into IFMIS systems, uh, when we start talking about the purchasing, because in public procurement, Generally, uh, we, we, we, the public procurement agency works with up until the contract award and then at the contract management it's handed off, uh, to the procurement entity itself, uh, many times, uh, to manage that and also deal with the invoicing, receiving, and payments. So we'll talk about, um, A couple tendering frauds, uh, collusive bidding, bid rigging, kickbacks, conflicts of interest, and purchasing frauds, um. Uh, such as, uh, uh, false inflated and duplicate invoices, shell companies, phantom vendors, and purchasing, uh, purchases for, uh, use, uh, or resale. Um, now we, we have, we've color coded the indicators, uh, to help get an idea of where they, they can be found and how they can be used. So the reds are, uh, real-time blocks or alerts for significant indicators. The browns are preprogrammed reports. Reports, uh, for other common procurement fraud schemes or waste or abuse. The orange are other less common reports to be listed on the handbook or online, and blue are links to public, uh, procurement records, telephones, and address information, which is very helpful. So let's talk about the first, um. Uh, the first report, uh, which we have, or the first scheme, which is bid rigging. Uh, bid rigging refers to the secret agreements between favored bidder and procurement personnel, uh, to improperly manipulate the procurement process to steer contract award to a favored bidder and exclude other bidders, often as a result of corruption. So there are many different common bid rigging schemes including change order abuse, uh, excluding qualified bidders, uh, leaking the bid information beforehand. The manipulation of bids, um, rigged or targeted specifications happens quite a bit, splitting up purchases, and you can, uh, identify these by taking a look at the sample indicators that we've identified, and these are all in the report, uh, and these are the primary data sources here that you would find. Um, each of these, uh, uh, of these various indicators, and algorithms can be created, uh, either through reports or BI or ex ante red flag indicators into your e-procurement system to identify, uh, this. The next one is collusive bidding, uh, and collusive bidding refers to agreements, uh, by contractors or suppliers to, uh, cooperate in the bidding process in order to avoid competition and inflate prices to artificially high levels. It can occur in small and large contracts where collusive bidding is well established, uh, prices can rise significantly. It's sometimes over 100%. Um, and there are various types of, uh, of this collusive bidding. One is complementary bids. These are, uh, can also be known as protective courtesy or shadow bids, um, uh, that are intended to merely give an, uh, appearance of a genuine bid, but it's not, uh, they're not, uh, meant to secure the buyer's acceptance or to win. There's also bid rotation, uh, where participants in bidding schemes or ongoing cartel activity often rotate winning bids on geographical areas or based on job type or uh timing. That's, that happens quite a bit. And then, uh, bid suppression where, uh, big breeding streams to succeed groups, members must prevent outside companies from bidding, and this can be done by either a payoff or, you know, through threats or potentially violence. And then there's market division, uh, the cooperating companies may divide markets, uh, and product lines and agree to not compete on each other's territory. So that all these happen quite a bit, and here we have simple indicators that can illustrate. Um, uh, it, it, by, um, measuring these, uh, these can provide, uh, help for risk, and then you can find these type of indicators in, uh, e-procurement system and IMIS systems, uh, under, uh, these types of data sources. And let's talk about, uh, our first, uh, purchasing fraud, uh, and, uh. So a contractor or supplier can. This inflated intent to defraud either acting alone or in collusion with a con. Uh contracting or personnel, um, and so you have, uh, false invoices, uh, which, um, our invoice information does not match the purchasing order receiving or payment information. Um, also there can be, um, uh, sequential, uh, invoice, uh, numbers and then inflated invoices where the invoice price. Quantities are greater than the purchasing order price, and the total payments are greater than the total invoice amounts and then duplicate invoices, uh, where the invoices with the same numbers, dates, and amounts, uh, are paid. So that's, uh, these are, uh, the different indicators that we can use, uh, to be able to identify, uh, this type of scheme. Um, additionally, we have shell companies. And with shell companies, uh, these are vendors referred to firms that are secretly owned by, uh, procurement or agency officials employed by the procurement, uh, agency. Such schemes, uh, such schemes, uh, are typically classified as conflict of interest under the general category of corruption. Um, And shell companies can also refer to fictitious companies set up by corrupt officials uh to act and portray suppliers and subcontractors in order to receive bribes and hidden assets. And here we can see the various indicators, uh, that you could identify, uh, uh, from these, uh, sources and include those in your algorithms, uh, for identifying shell companies. And then we also have, uh, phantom. Vendors and Phantom vendors are, uh, are our procurement or payment personnel that can create a fictitious, uh, contractor, consultant, or vendor or supplier that does not provide any goods or services in order to embezzle project funds. These fictitious companies often provide consulting. And other hard to verify services such as a repair work or deliver consumables rather than tangible goods and services that can be uh later verified. Um, and then we also have the resale, uh, or purchase of personal use and resale, um, or uh diversion. Um. Some indicate sample indicate here, uh, uh, you can have a, a, a, a different ship to address, a high number of purchases of, uh, certain items susceptible to personal use such as fuel or laptops, uh, or gas, uh, unexplained, uh, spike in purchase of such items, and you can get this information, uh, from these data sources here from the the vendor and product codes, PO invoicing, shipping information records for. Um, So, uh, a lot of these also have, uh, to, uh, one way of, uh, not validating, but identifying risks is by applying Benford's law, which allows you to, um, to take a, uh, Benford's law works with, um, any set of numbers generally. Uh, take the, the populations of all of the countries in the world, if you put those on a piece of paper, you would find that, um, That the number 1, would be there about 30% of the time, uh, the number 2 would be there about 18% of the time, and the number 3, it kind of goes down in a logarithmic fashion. It's uh an interesting, um. Uh, uh, fraud detection, uh, resource that can be used, obviously, uh, and it can help to indicate fabricated numbers, uh, and fraud. And, uh, again, uh, you would need to validate this, uh, before you, uh, move forward with it, but it's a definitely an, a way of identifying, uh, a risk. Um, there are many, uh, commercial products on, uh, the market for detection of ifus, where in e-procurement, uh, we don't have, uh, there, there are some, um, but they don't, um, they aren't fully integrated in with e-procurement systems as of yet. Um, and they, lots of times they only go into the contract award. Um, whereas for, uh, IMIS systems there are, uh, modules on the market that you can, uh, purchase and install. Uh, in fact, um, So, uh, there are a number of robust, uh, commercial fraud detection prevention systems that can be installed or linked, uh, to IFMIS and ERP such as SAP, um, and these systems can provide continuous monitoring and ex-ante alerts, uh, or provide, uh, fraud, uh, uh, of potential fraud, many related to accounts payable transactions. Um, so here we have, uh, the SAP HANA fraud detection and you can see many of the features that this provides. Um, more than half of ITMIS systems, uh, funded by the World Bank or in, uh, in developing countries do use these, uh, um, commercial products, uh, and then, then others can be, uh, homegrown. Um, this particular SAPPANA fraud management. I programmed to detect and investigate and prevent fraud in day to day processes including uh order cash, uh, procure to pay and product, uh, and planned a product, request a service, and core capabilities. Another product out on the market uh is Appen, and it's an AI learning, uh, a machine learning audit and spend management application that links to IFMMIS and other business management systems. Um, it's similar to the SAP module, uh, and it can identify anomalies and possible fraud in accounts payable. Um, one interesting thing with apps in is that you have, uh, they have, uh, the ability to use AI technology to read and interpret written content in documents to identify anomalies and compliance issues. Um, and, uh, and it can be used to read the text of receipts, uh, and disallowed items. We also have Oversight Systems has a lot, and, uh, Galvanize has, uh, ACL, uh, Essentials is another product, uh, on business. Um, what other drivers could be, should be considered? Well, other drivers, uh, is we need to make sure we have the data available, uh, so that, uh, it can be used. The data needs to be available from. Um, uh, from the IMMIS and, uh, procurement systems, uh, and these also, there should be data privacy, uh, and protections around this data so that the folks that are only allowed to view the data can view it. Um, we also need to make sure that the, the systems are interoperable, uh, and that, uh, we're integrating with, uh, existing, uh, detect. Um, so we will, I think the big thing that we found with the study is that, uh, This Hunt we're having difficulty hearing you. And you have, you, you are breaking in and out. Let me see if I, OK, is that any better right there? Mhm. A little better. Yes, OK, OK, very good. Um, let me Let me get to the right page here. OK, what are, uh So Uh, application benefits. Uh, detecting fraud and corruption, particularly in procurement, is a major benefit. Most current procurement, uh, fraud detection algorithms are expost, and we would prefer these to be ex ante. Uh, and e- procurement systems can certainly help with that. Um, for, uh, anti-fraud systems that can be installed or linked to commercial ITMMIS systems to improve continuous monitoring and ex-ante alerts, um, and similar, uh, functions can be programmed in homegrown systems. So our key recommendations that we have is we want to ensure, um, the importance of moving to ex ante fraud detection. Uh, we also need, uh, need to ensure that we're using a stronger fraud detection algorithms, uh, that, uh, traverse multiple systems such as e-procurement and IFMIS. Um, and the, uh, fraud detection algorithm should be tailored to count, uh, to countries where they are installed. Um, the fraud detection technology is most effective when it's integrated in with traditional detection and prevention methods. Um, and, and should, uh, the automated fraud detection should be extended to project implementation stage. Um, And I do want to discuss real quickly what the next steps, what we think would be uh nice to do is create a prototype for ex ante digital fraud detection program for procurement and put and installing fraud detection. Systems to run remotely and independent oversight organizations. Also, I think it's very important to include fraud detection systems into gov tech projects where there is political will and uh digital infrastructure capacity and, uh, enthusiasm. So thank you very much. I, I look forward to IS's presentation. All the best. Thanks, Han. Let's move uh swiftly to, to ISA and to hear about the fraud detection in HR systems. Over to you, Isa. Thank you so much, Tracy and Han. So let me, uh, connect my screen. And so I hope that is working and you can see it right now. Um, OK. So, let me begin. Thank you so much, Han, for your presentation on the IFMIS system, um, And thank you so much, Ed and Tracy, for your initial comments as well. Um, I wanted to start off a little bit by talking about, um, you know, corruption, good governance, digital technology, sort of setting the stage, then take you to a more focused discussion on HR systems and how we can apply digital tools to HR systems to help detect and prevent fraud and corruption. And then I also want to take some time towards the end of my presentation to talk about the analog components. Um, Hunt has mentioned them in. Very effective fashion already. Um, so it doesn't bear repeating, but I will highlight some of the particular analog compliments as they relate to HR systems. Um, so, moving on to the first section, which is kind of looking at corruption, good governance, And digital technologies. Um, can, can you show us the presentation? We cannot see from our side. You can see it, my apologies. Let me Let me do it again. Please let me know if this works and apologies for the hiccup. Does this work now? Yes, can you make the screen? Yes, OK, it works well. There we go. OK. Month 11 of, um, quarantine and we still haven't figured it out apparently. OK, I think so now with the hiccups past us, let me begin again. Um, so, as I said, this part of the presentation will focus on HR systems specifically, and how we can use digital tools to help fight fraud and corruption in the public sector. And towards the end, I would also reflect a little bit on the analog complements to digital reform. So, kind of setting the stage, um, I think Edge and Tracy in their comments really talked about the importance of digital tools to fight corruption. Uh, we find that there is a very intrinsic link between Um, you know, corruption perceptions and the ideas that we have regarding good governance, these are measured very effectively in global indices as well. So, as you see on the charts over here, on the left and the right, on the left, we have the worldwide governance indicators and sort of these pillars that we consider to be, um, you know, Good governance or ideas like rule of law, regulatory quality, voice and accountability, govern government effectiveness. Um, and when we find, uh, what we find in our data analysis, and this is taken from a very good publication that was produced by the World Bank, um, is that control of corruption is very closely tied to a lot of other. Um, indices related to good governance. So, a country's performance on many elements of good governance are closely tied towards its performance on the control of corruption. And the control of corruption pillar is also very closely tied with how citizens perceive their government and how much trust they place in their political actors. And that's what you're seeing on the chart on the right-hand side. Over here. Um, what it shows essentially, if you look at sort of the, the, the scale here, as you move from 0 to 100 in terms of perceptions of corruption, 100 indicates that, uh, people perceive corruption, the government to be absolutely clean. So, there are no instances of corruption that are taking place. And on the left hand side, what you're seeing is that as you go up from 0 to 7, Has greater public trust in politicians. And here, although this is not a causal link, it, it does very much show that there is a close correlation between these two elements, which is to say that as this perception of corruption, um, decreases, as more and more people think that government is clean, that officials are not corrupt, they are. And greater and greater trust in public officials as well. So, it's very important for governance, um, measurements and governance outcomes. At the same time, what we see is that there hasn't been much improvement, right? So, if you look at, again, the WGI's index on control of corruption, and you compare what's been going on over the past decades, You see here for the majority of the regions, there hasn't been much progress. In fact, percentile ranking has actually declined in regions such as LAC, Middle East, South Asia, Sub-Saharan Africa, even in reporting that we see in other measures such as Transparency International's, uh, Corruption Perceptions Index, what you find, um, is that a significant number of countries have been performing very poorly year on year when it comes to corruption outcomes. Um, so with that, what are the opportunities that digital technologies can present, right? Um, here in this slide, what we're trying to demonstrate and what we were able to find. In the course of working on this report, is that um a lot of digital technologies are creating new frontiers for the fight against fraud and corruption. So, you see some examples here of governments taking direct interventions to fight corruption, like in Albania, where the government created an SMS-based platform for citizens. To be able to report, uh, when they experience somebody asking for a bribe from them in order to access a government service, or when they witness somebody else giving a bribe or being asked to take a bribe. And this really helped elevate the government's fight against corruption. They were able to increase the number of corruption investigations that they were taking on. And led to positive outcomes. Um, on the other hand, and this is something that Ed alluded to towards the beginning of our discussion, is that there are a lot of indirect benefits to digitization as well. So, as governments, uh, take on this digitization effort, as they digitize more and more government functions, what you find. Happening is that the opportunities to engage in corrupt behavior and fraudulent activities actually closes down. And so there's this indirect kind of an externality, um, which is very positive in that there is less corruption taking place. And that's the example that we see here, um, in Afghanistan, where, The creation of this M-Pesa mobile payment system allowed the government to stem a lot of this stiffening that was taking place in terms of salary payments. Um, it was actually very interesting. Employees thought that they were receiving a bonus. Their incomes were going up by an average of 10%, when in fact what was happening was that that 10% was being lost, um, in the process of them receiving their cash income because it was being taken up by either middlemen or ghost workers, etc. So there was this overall benefit, and we're able to capture it in data as well, which is what you can see on the right-hand side over here, which is, again, not a causal link, but a very close correlation, which shows that as you um promote, as you expand more and more e-services, more and more e-Gov functions, uh, And perceptions of government actually start to decline. So again, the scale on the bottom, 0 to 100, with 100 indicating clean government, and on the left, 0 to 11 indicating, um, you know, this full development of e-services according to this UN index, and this very close, tight relationship between the two taking place. So, moving to sort of the public sector then, I mean, we've talked about how anti-corruption efforts are very important for good governance outcomes. We've talked about how digit technologies offer a very exciting new frontier. But what we find is taking place within the public sector is that there's still a lot more room for improvement. Of course, this also means that every year there's new and greater innovation. So, every year, there is more and more that governments can do. Um, analysis that we undertook, however, showed that within the public sector, there is still an increasing reliance on very traditional mechanisms such as sort of audits, um, relying on tips from whistleblowers to tackle fraud and corruption. And we think that much more can be done. And in fact, when you look at the application of digital tools, what you find is, um, that you can run a lot of algorithms, you can create a lot of programs that can be much more effective in identifying patterns of corruption, in identifying risks before. And actual fraudulent activity occurs. So that's sort of gonna be the focus of my presentation, that's the focus of our presentation as well. Hunt talked about IFMAIS systems, procurement systems. Um, I wanted to give you kind of an overview of HR systems and how they can be deployed. A very important point to make here is that whereas traditionally IMIS systems and procurement systems have been used in the past, um, for in terms of government efforts to identify fraud and corruption, HR systems have not been similarly utilized. Now, this makes a little sense because when you're looking at large procurement. contracts, when you're looking at IMS systems that are tied to government's budgeting, uh, there's this dollar value that you can assign, right? So there's a dollar value that you can assign to a fraudulent activity, to an instance of corruption, which are, whereas HR might not present a similar cost to the government. But the point that we make in our report is that any instance of somebody breaking the rules when it comes to HR management, be it attendance, performance, etc. Uh, it represents a loss to the government. And so that's something that we're going to talk about in the examples that we share, uh, which is either the loss in terms of money, whereas that you're paying a wage to an employee who's not performing, so it's a loss of money, but also the act of that person. not doing their job effectively represents a loss of performance as well, a loss of government outcomes. Um, at the end of the day, the core role that the government performs in any society is to provide services to citizens, is to generate some kind of outcomes at the end of the day. So, um, before I jump into sort of very specific examples and share them with you, I wanted to just provide a very quick overview of, um, an HR system and what it typically looks like. In any country. Um, now, this represents sort of the ideal, it can vary a lot country to country, but, um, a lot of countries do have some kind of HR system, right? Um, and this HR system will typically capture information regarding the public sector employees' biographical information, their attendance records. Records, information about, you know, the job position, title, what department they're working in, um, how much salary they're getting in form of basic pay or allowances, etc. and typically these kinds of HR systems are also linked to the government's budgeting system such as IFMIS or a procurement system, so that you can really use these systems to have an overall view of government as well as use it for policymaking and public sector management, um. So, the right-hand side is just a snapshot of South Korea's HR system. It kind of represents the idea. Um, it's meant to just give you an overview of the different elements that such a system can include, um, as I mentioned, sort of biographical information, attendance records, it could be linked to performance evaluation, training, development, etc. as well. So let me jump into examples. What I'm going to do is give you three specific examples of um the kind of fraudulent activities that can take place in the public sector that are linked to HR and how an HR system can effectively be used to tackle them. The first example is related to employment and attendance records. The example, um, the examples that I'm going to share here with you kind of illustrate what the core challenge is, right? And the core challenge that we often see in public sectors around the world, is that you might have a lot of ghost workers, so employees that don't exist, in fact, but are actually on the government's role. They're collecting salaries, but they're not actually showing up to do any work. And then there's also employees who do exist, um, but they tend to not show up to work at all for long periods of time, with unexcused absences, etc. An example from India is just one such illustration, um, where the study from 2016 showed that on any given day, Up to 25% of government teachers and up to 33% of healthcare workers did not show up to work without a legitimate reason. So it's a very significant problem, not small at all. It does represent a very big cost in terms of wages that are being paid, in terms of work effort that's not being undertaken. Um, and there's several digital solutions that can be applied here. The first that we've shared here is, is something that's already in use. So digital ID systems, a lot of us in the audience might be very familiar with this. Um, their use is expanding across the world. The example here that we've shown is from New Guinea where they were able to create a biometric ID system for employees. It was linked to the HRMIS and in the process of kind of ensuring through biometric registration IDs of employees, they were able to eliminate a lot of fictitious positions and ended up saving, um, Up to $1.7 million US dollars, which is a very significant amount. The other example which we're also seeing in more and more countries around the world, is using biometric scanners. Um, so, you know, an employee shows up to work, they scan their thumbprint, they go into the office, they do their work, they leave at the end of the day, they scan their thumbprint again. It might not be extremely effective because somebody could show up to work and just sit and not do anything, but it is one way to ensure that attendance is taking place, right? And so what you can do is track those attendance records, link them to the HRMS, make sure that you have um an algorithm that's doing iterative tests, and it can raise a red flag when somebody is chronically absent, not showing, Up to regular working hours, etc. So, this is an example where you're seeing a lot of these biometric scanners in government buildings, uh, but perhaps they're not being utilized most effectively because they're not being linked to the HRMS and not being run through these kinds of tests, which will be very helpful in actually tracking where there is a lack of apps, a lack of attendance by government employees. The second example that I wanted to share with you. Um, is in the case of recruitment and promotion decisions. So, here again, a very common public sector challenge is that, um, a candidate might submit false records, um, in terms of, you know, for the hiring, um, as well on the other side, you might have hiring managers who tend to favor certain candidates because they're, You know, maybe tied to them through family links, or maybe they're just expressing some kind of favoritism and impart impartiality to one candidate over the other. Um, and this can be a very significant challenge as well because at the end of the day, what you're taught, what you want to promote in a public sector are meritocratic principles, right? So here there are a lot of digital solutions that can be applied to the HR system that you have with regards to remote recruitment and promotion decisions, um, to make sure that this kind of bias and this kind of sort of fraud and colluding does not take place. Um, some examples that we've shared here in the report and that I wanted to highlight here as well, um, are things like, you know, connecting the, um, HR registry with common registries like academic. Registries, police reports, etc. and you can run these tests, tests and flag where, um, you know, a candidate that you're hiring actually has, you know, you can verify their academic credentials, you can do a background check and see if they have any criminal history that might pose a problem for their employment status. Um, similarly, if you're trying to hire somebody who's an internal candidate, you can run these kinds of checks in their HR records to show that. In the position that they're potentially going to be hired to, um, you know, they have the necessary qualifications, the years of experience, the academic records, the sort of professional skill set, etc. Uh, similar cross-referencing of data elements that you can undertake is to flag if one single individual has been responsible for hiring a, a, a very significant number of people, right? So, if somebody is working at some ministry, uh, in local government, and they're just hiring a lot of people to favoritism, to nepotism, um, if a lot of that sort of is originating from some, a small subset of individuals, this is a common test that you can apply to kind of flag and explore further, whether this is problematic. Or if this is just in the course of significant hiring that needs to take place and the individual is just performing their duty. Um, another common example, uh, that can be applied is that you could have, um, an algorithm that kind of tests for very common last names and family names. Um, of course, this also requires some kind of human oversight because some family names are very common, uh, but it is a potential. link between uh families that could indicate somebody's hiring because they are related to the candidate. Um, and so the system could flag these and then, um, you know, actual human auditors could come in and verify what is going on in the situation. So there's lots of different ways in which really you can apply, essentially kind of an add-on to your existing HR system. That helps run these tests, helps raises these red flags of potential problems with regards to recruitment decisions and promotion decisions, and then you can use human oversight to actually um investigate what's going on and stem a problem that might be occurring. The third and final example that I wanted to share with you, um, as an illustration of how HR systems can be used to detect fraud and corruption. As in the case of employees' salaries and income. And so now here, a very big challenge that you see in public sectors around the world, um, is that you have government officials who take bribes, right? So this is a very common problem reported by citizens all over the world, in order to access a very basic service like um getting electricity, uh, being able to see a doctor, get medicine, or enroll their child into school, they're asked for a bribe by a government worker, uh. Similarly, you have government workers who might be double dipping, so this is the idea that an employee is actually, um, pocketing an allowance, an additional benefit, an initial bonus, or more, a double salary, which you're not actually eligible for, um, and typically what's being done right now is that you'll have some kind of auditing, some traditional audits to try to verify the salary payments. Um, but again, this is also an avenue where you could use digital tools instead, um. To kind of stem this problem before it occurs. So some of the things that you could do is to try again to cross-reference the different data registries and you could say, well, this individual, you could look at their bank transactions, their tax filings, other income filing, disclosure documents, and see if there are any suspicious transactions that are taking place, any transactions taking place. Is above a certain currency level, that might indicate that there's something suspicious going on in terms of their actual take-home income, which might be out of line with their core job within the government sector and the salary that they should be getting typically from that government job. Um, this is also something that you can do for your existing employees, right? So, You can look at an employee, um, this person in your HR system is noted for this position title, working in this grade in this ministry, and this is the actual salary payment that was given out this month to this person. And if that is out of line, if the system is indicating that this person according to the grade and title should actually be eligible for only this much amount or. They should be eligible for XYZ allowance, but they're also getting two other allowances which they're not technically eligible for. The system could flag these and then you could go in and actually make sure that that system is cleaned up and that employee no more receives those salary payments or those allowance payments. Um, and then finally some of the other things that we found that could be very easily applied to HR systems. Are things that, um, you know, you could just have this system running these iterative tests and you could see what is somebody, um, has somebody indicated that their home address is actually a PO box or a mailbox? Are they receiving multiple paychecks? Are they getting a paycheck or a bonus, um, that's out of line with sort of typically when this category of payment is paid. Out and the system could very well generate that. These are the kinds of things that are very, um, would take a lot of time and effort if a human being had to do it, but if you design the system to do it, it can be done very cost effectively, very easily, and the system might be able to identify many more such instances than a person could. And this could also be applied to things like travel expense reports, overtime checks, etc. uh, really just any category of payments that are being paid out to workers, um, where the system could tag these very effectively. So that's sort of, um, kind of examples, there are a lot more that could be applied. Um, and again, what you'll see here, um, if you go through the report, we've provided country case studies where they are available, um, but a lot of this is hypothetical because it's not being applied to the HR sector yet. However, given the state of digital technologies right now, there's a lot more that can be done, even in the HR sector, and I think it's important to realize that. You know, when you're looking at big bidding, procurement, if the systems, those are important. But these kinds of activities where somebody's not showing up to work, collecting salaries and payments that they're not eligible for, um, when they're lying, say, or providing false records for recruitment and promotion, these are all very important dementias as well that the government should address and they affect the overall functioning of the public sector. So, um, just the last bit that I want to spend maybe a couple of minutes on because I'm mindful of the time as well. Are some of the analog complements to digital reform, um, Hunt kind of delved into these, so I don't want to take up too much time, but I wanted to just quickly flag some, they're also very relevant for HR systems, not just procurement systems, right? So, data availability is a very big one, We talked a lot about HR systems and cross-reference in data registries, but if the data doesn't exist, the government doesn't have an ability to do any of that. Um, and in fact, what we find a lot of the public sectors that we actually work with, um, as part of our work with the World Bank is that often public sectors do not have very comprehensive, accurate, timely, centralized data systems. So certainly that's something. To begin working towards and then also in parallel, you can develop these digital systems to detect and prevent fraud and corruption. Data protection, of course, is also very important and Hunt alluded to this. Um, you know, individuals have an intrinsic right to data privacy, and therefore, the government has to really exert an effort to make sure that they identify, you know, where, uh, an employee's. Data will be used for what purposes, by whom. You have to make sure that there is effective cybersecurity, um, you know, protocols to prevent these threats, prevent hacking instances and cyberattacks, um, make sure that you have the digital tools that you're applying, but that government employees then don't misuse these digital tools to go after their own opponents. And then also to make sure that there aren't inherent biases that are creeping into your algorithms and data protocols. And this is where you have to constantly retrain your algorithms and your digital tools. Um, and then of course, there's data infrastructure and data sharing, Han talked about this, the public sectors that we work in, um, the culture is one of working in silos, not really sharing information, not really working together, and so when you start talking, About these data registries that need to be cross-referenced, um, across government, across sort of outside of government as well. What you really have to come up with are these data sharing protocols, you have to assign roles, make sure that business processes in place. Um, again, who shares data with whom, for what purpose, in what manner, and how that will overall help them in their fight against, uh, corruption and fraud, um, data. Literacy, digital literacy, and institutional capabilities again then come up. Um, do you have the right skill set within your public sector for your staff to be able to use these digital tools effectively? And also, are you creating a pipeline that you in the future, have such digitally savvy employees in place? And this would really require working with academic institutions and making sure that the local curriculum is addressing that. And then finally adapting to the local context, um. Again, this is something Hunt talked about in relation to procurement and IMM systems, um, similarly relevant for HR systems as well. Um, what we found in some of the analysis that we did was that, um, you know, where you already know that there is risk of corruption, where you already know that this is typically where, you know, and this is what we've done, so, ghost workers, somebody siphoning off like additional cash allowances when they don't, they aren't eligible for that, typically. In any government around the world, people have an understanding of what the risks are for fraud and corruption activities. And so you should design a system that adapts to that and helps you track that, um, instead of just importing sort of this comprehensive digital system that might not address the root causes of the problem as they exist in your local context. And finally, linking to government decision making is intrinsic, right? So you could have this very fancy digital system, you could have all kinds of. Algorithms in place, um, you could be using it and you could be tracking it, but if nobody's taking any action on these reports, um, if nobody's sanctioning employees as a result of their fraudulent activities, nothing will get solved. And so here you might have to develop certain regulations, um, and civil service procedures to make sure that employees can in fact be sanctioned for their fraudulent activities. Um, so that's sort of it for my presentation. Thank you, and let me stop here. Thank you, Isa, and uh just to, to thank both presenters uh for their excellent uh coverage of the issues and maybe to also make an observation that the report is rich in these case studies and these examples that I think are really uh helping to drive home some of these key messages. Um, one of the points. I just made at the end reminded me of the example of Brazil where the Comptroller General introduced AI to detect who was most likely to be a corrupt civil servant, but the law hasn't adjusted to allow such things to be evidenced. So please do read the report because these examples really illustrate some of the points being made. Let's turn now to our discussions. Uh, first turn to Donna Andrews, who's a senior public sector specialist, and, uh, then we'll come to Helena Amaniova, uh, who's a senior operations analyst for their key points before coming back to the Q&A. Donna, the floor is yours. OK, um, thank you, Tracy. Um, and first, um, thanks to Hunt and Izzo for some really interesting presentations. There's, um, plenty of information and food for thought in there. And I must say that this topic um I guess takes me back to a life before the World Bank, um, where I was actually the head of an ethical standards unit in a public service commission in Australia. And so I know that the issue of managing fraud from staff is a universal challenge regardless of where you are in the world. So I'm certainly glad that the bank is making a contribution to and challenging the ways in which governments can tackle this issue. Um, I should probably also, um, I guess confess that I come from a HR background, so I'm, I guess somewhat biased about the importance of, of HRM and the contribution that it can make to government effectiveness. And so I'm going to focus, um, my comments largely on, um, ISA's HRM, um, presentation. Now I think the, the timing of looking more closely at how technology can help to identify HR fraud really couldn't come at a better time. So in most of our client countries, um, wage bill payments form a significant part of government. Expenditure and the ongoing impact of COVID is going to mean that the government expenditure is likely to be under more pressure than ever. So it's, it's really critical that every dollar is spent well and that it gets to its intended destination. And the first thing I wanted to comment on and perhaps this speaks more I guess to my HR background is um and and Isa has talked about this, is about the underutilization of HR data. Um, you know, HR captures a huge volume of data about employees, so everything from um date of birth, family members, national ID numbers, remuneration, qualifications, banking information, really the, the list goes on, it's a huge volume. So there is um a requirement to collect it and to store it, but generally then it's not referred to again unless possibly um a problem arises somewhere else. So in comparison to the procurement and financial information, there really isn't a strong history of using technology and data from HRM systems to help find fraud and misconduct. In fact, I I don't think I'm, I'm too out of line to say that there really isn't a strong history of using data from HRM systems, even to inform decision making and planning more broadly. So it's definitely an area, um, particularly where HRM professionals can and and need to do better. The issue of um technology and data analysis as key competencies for HRM professionals, um you'll find is a key feature of um competency frameworks in almost all advanced economies and most OECD countries. So there's a recognition that we need to do more and to help HR professionals to understand, um some of the the key skills and experience that they need in that area. But I think that um ISA provided some great examples, but a stronger focus on using data from HR systems can certainly help to avoid some very costly fraud situations, um and particularly I think the example that that springs to my mind is, um recruiting individuals without the necessary, Qualifications. Um, I can recall a number of specific examples from my time in government in Australia, where staff members were recruited to positions which required specific qualifications, only to discover at a later time that the individual didn't have those qualifications. Um, and I, I can recall an example um of a doctor who was in fact recruited to a regional hospital. And the issue about the qualifications only arose after a series of complaints about poor patient care and outcomes. And I think in times of COVID, ensuring that appropriately qualified staff are being engaged for vaccine programs, for example, it seems like a really critical decision and one where using technology to validate qualifications could certainly save significant time and money. And the US Department of Labor recently estimated that the cost of poor recruitment decisions can be up to 30% of an individual's first year salary. Now that can be a significant amount of money, um, and I think that that that money could certainly be better invested in perhaps learning and development for staff or even um improving IT systems. So I wanted to um just add a couple of further examples in terms of recruitment and selection um on top of what ISA has has talked about already, where technology can provide some great opportunities for for fraud identification. So I guess while technology has brought some great opportunities to streamline, um, sooner we can now use AI to screen um CVs and job applications, we can use one-way interviews where candidates can respond to pre-recorded interview questions during a timed online interview. Uh, we can use chatbots to um conduct initial screening questionnaires with candidates. But it's also brought the opportunity for some candidates to use technology to try to gain an advantage or seek to deceive recruiters. Um, so particularly in the private sector, organizations are adopting some, some high-tech solutions to help better manage the risks. So for example, um, looking at using blockchain technology to identify employees, um, individuals who have, um, provided fake qualifications, uh, using facial recognition in video interviews to ensure that the applicant is actually the one who is being interviewed. Assessing voice and speech tone and using some sophisticated written assessment tools to ensure the the authenticity of candidates claimed skills, so there is a huge volume of these and of course they're evolving and moving forward um um continually. The third issue I wanted to touch on just briefly is the analog complements that that Isa Anne Hunt referred to. Um, for me, I think one of the key takeaways is really the the confirmation that the digital tools aren't a be all and end all just for themselves. We really need to combine the digital reforms with the analog complements in order to be able to make the most out of the technology. Um, and I think one issue which is worth underscoring, um, again that that is a raised is about the staff skills for technology. So the effective use, being able to effectively use HR data requires staff who are familiar with the HR systems, but also the type of data that's available, but also that they've got the data analysis skills in order to make sense of the information and to be able to use that as part of identifying potential problems or issues. So traditional fraud and misconduct investigations um definitely still rely on the human dimension, but technology can play a great part in helping to point us to areas where we should look deeper, well before we, we might otherwise know that a problem exists. Now we know within, within the World Bank from surveys of civil servants that many governments face some basic constraints, including a lack of digital skills. Um, for example, in the Philippines, which is a middle-income country with a fairly vibrant digital economy, um, only 8 out of 10 staff could in fact create a PowerPoint presentation. And if we look at um surveys from Ethiopia, in fact, fewer than 50% of staff in government could use a computer to write a memo or to create an Excel spreadsheet. So given the speed at which technology is moving, there certainly needs to be continual focus on staff skill development. And in fact within the bank we're currently working on a report which will focus on gov tech skills in the civil service and how to build and and retain those skills. So finally, um, let me um thank the team for a great and really practical report and some excellent examples of, of how technology can help, um, with fraud detection in public administration. I certainly encourage you all to, to download and to read that. And perhaps let me just um finish um with um a quick question for ISA, um, which hopefully we might have some time to to look at is, given that particularly in HRM the use of data and technology is really in its infancy in any of the countries that we work, um, what are your suggestions about how countries could get started on this type of agenda, particularly in um the HR area? So thanks very much. Thanks, Donna. Let's move quickly to Irina who's been very patient, uh, waiting for her turn. Irina, over to you. You're the sound quality is not working. Elena, can you um try to mute and then open or? Microphone again Oh, we, we cannot hear you. Um, Actually, I suggest that you can you can you um turn off the the um the blur, the blur effect. It might help So why don't we let Irina and Natalie solve the audio problem and whilst we're doing that, um, I will pick up a couple of the questions, um, back to the presenters. Um, Natalie, maybe you could read Irina for a minute. And so the first question back to the team, uh, what, what's the actual evidence on the financial management system's improvements when it comes to issues of fraud and corruption? A question from our colleague in Malawi says that there's a lot of perception, but in fact, Incidences of fraud in payments seems to have increased with the implementation of online financial management systems. Maybe we could turn just to that question, Hunt, and there's another question in the chat for you if Irina doesn't come back online, which is how is potential fraud monitored during project implementation. Over to you, Hunts. Great, thank you so much. Now, um, one of the, uh, regarding, uh, fraud and payment systems and uh the detection of digital detection of fraud and payment systems, um, in the Uh, in the countries where we work, uh, um, in, in IMIS systems, very rarely are these fraud detection systems, uh, installed and used. Many times there is a political economy issue with, uh, having, um, Uh, getting them installed. Uh, so, and additionally, uh, a capacity issue with getting them up and, uh, running and, and monitoring, monitoring them. So, I'm not aware at the moment of any, uh, implementation of these modules in, uh, ITMIS systems. Um, I'm more of a procurement expert. Uh, uh, this is something we could ask, uh, some of our IMIS experts, uh, at the bank. Um, but they, they are available, uh, and they are certainly used in OECD countries and, uh, in the private, uh, sector also. Um, so thank you for that question, and I, I think there, there is a huge, uh, uh. Uh, a huge, um, possibility or opportunity, if you will, to uh incorporate some of these, especially in ITMS, uh, uh, some of these, um, uh, fraud detection, uh, modules. Um, in, in e-procurement, for example, what happens is, is that, you know, Countries are, are so focused on just digitizing the paper-based e-procurement system that, you know, they're, they're not looking for a fraud detection system. They're looking for an e-procurement system which helps them digitize the procurement process. So, um, we're, you know, we're thinking of, of, of enhancing uh the uh the, the e-procurement. Um, I guess, uh, projects that, you know, phase phasing it out, uh, and adding an integrity module, uh, and later phases of a new procurement system, um, regarding the, uh, how do we, how do we monitor, uh, fraud detection, uh, in, in, in, uh, bank projects? Well, it's interesting, um. Uh, it was very, it was, uh, before the new procurement framework, uh, was adopted in, uh, July 1st, 2016. It was per transaction, right? If, if a procurement specialist who was doing a prior review saw something, some hanky panky, uh, in the bids during the evaluation process and the evaluation report, then, uh, we would, we would take a look at it and if it looked as if, uh, there could be some fraud there, we would send it to INT and INT would do the due diligence of following up on that. So that was historically how it was done and that's how it's done now on an individual basis, um. Now that we have SEP, we have a lot of useful information in SEP, and I know that our colleagues in INT and ISP, uh, are, are building a monitoring, uh, uh, algorithms that can look at this data. Now, SEP is really just a, a, a procurement monitoring system. It is not, uh, an e-procurement system, but there's a lot of very useful information in there to help, um. Uh, help, uh, do that. So that would be my, at the moment, we do it on an individual transactional basis and still somewhat manual. Hopefully in the near future it will become, uh, much more aggregate, uh, with other fraud detections. Thank you. Thanks, Hunt. Irina, let's, uh, let's try again to, to connect you. We, we, we still cannot hear you. Can you try turning off the video? That might help. Can you try um your audio? Um, OK. OK, Irina, I'm afraid we still don't hear you well, um, unfortunately, I know that this did work in the, the, the beginning of the session, so something happened over time. Maybe Irina, you could, um, reconnect through another device, um, and maybe leave this one on for video and use a different device for audio. So, uh, turning to the next question in the chat, I'm going to paraphrase it a little bit, um, but as we use more sophisticated tools to detect wrongdoing or bad behavior, how long is it before the people who would like to circumvent these, what is it, they're circumventing manual processes now, get smart enough to circumvent the e-tools or the new IT systems that you're bringing in. So how long does it Take before somebody knows how to go into the system and, and just talk to the, you know, the qualifications, um, or to, you know, ensure that the red flag isn't shown up in the procurement system. So question back to you, is it a hunt as to whether or not this, uh, whether or not the government can always stay one step ahead or not. I said, do you want to go first? Sure, let me go first and then, um, Hunt, please feel free to add on to what I say. I mean, this is a very good point, right? And, um, I think the fight against perhaps fraudulent behavior will be. Always there, uh, as much as we might not want to see it, uh, but I think that's why the report focuses not only on describing what these digital tools can look like and what the system can look like, but also emphasizes the analog complements. Um, and this is really where you can try to make sure that as people then try to game the digital tools themselves, the government has systems in place to make sure, um, that that does not happen. So things, for example, like, You know, uh, talk, we talked about digital literacy, but also promoting a culture that is ethical. I mean, it's not just about digital systems that help you catch somebody after they've done something. It's also to actually create and promote a very ethical culture within the public sector where you have guidelines that say, look, like this is what we stand for, these are our values, this is what we don't stand for, and this is why behavior that is unethical is not only bad for business. But it's also something that we don't believe in as a, as public servants. So, I think that culture also plays a very important role on its own, and government should put in place, um, you know, training modules, guidelines, value systems, trainings that can help promote that. But in parallel, of course, you should also develop systems that help you stem these kinds of unethical behaviors as they take place. Um, and this is where I would like to refer to something that Donna mentioned, which is a very good point. Which is that, again, digital tools are not a panacea. You have to pair it with human intervention. So, um, it's about having still the auditors in place, having still the investigators in place, having still the regulators in place, um, making sure that they use technology as a tool that helps them address fraud and corruption much more effectively, but still, Having that human in place who can try to stem the tide and kind of be one step ahead of people who might still try to misuse the system. So, I think for me, the idea is use technology as a tool, still have the human intervention, and then make sure that you're also working towards expanding the ethical culture within the public sector. But, um, Hunt, any additional points that you would like to add? Uh, I, I would, I would say, you know, uh, it's a great point because in the future there's gonna be different types of fraud, but let's focus on the, the, the ones we've identified in this report, we have, uh, ISA has, has presented, uh, an, an amazing set of, um, schemes, uh, and algorithms. Um, that can identify, uh, fraud that's happening right now. We also have this for IMS, and we have it very well spelled out for procurement. So focusing on those now can really enhance, and it's, and it's using the digital side of this, uh, with an aggregate, uh, look at what's happening. Um, and, and I think that the speed of once it becomes all digitized, uh, it, it, it, it, it's gonna happen pretty quick. Um, there will be new fraud schemes, they always come up, um, but I think the speed of attacking those is gonna happen much faster once everything becomes digitized. Thanks. Thanks, Susa, thanks, Hunt. Irina, let's, uh, let's turn to you. Ah, thank you. So, now you can hear me, I hope. Thank you very much. Uh, so first of all, uh, thanks a lot for this great presentation and for the information that you provided to us. Uh, I would like to focus on fraud in procurement because I have a procurement background and now I work in OPCS anti-corruption program. Which is also analyzing fraud and corruption cases in our operations, and it is very, really important to identify fraud. At the earliest stage, so the tools which propose ex ante ex ante measures to identify fraud is very important indeed. For instance, our analysis of fraud and corruption cases in operations shows that fraud cases are just Fraud at the bidding stage is a prevailing type of fraud and corruption practices, and it occurs either on its own or in combination with other fraud and corruption practices. And ours, if they are trained to detect this fraud at the billing stage, they We can effectively identify it and prevent fraud and corruption in our projects and in our contracts, but unfortunately, according to our information, the can identify fraud in our operations only about 50% of cases. Training and also sharing the information on on digital tools which can prevent fraud and corruption is very important and maybe we can also think how we can use this information which is presented today and which is described in. The report. How can we use this information to share, to share the tools, the options which exist, but also to provide more information on governance filters that you can apply to detect fraud and corruption? How can we share this with our operations with procurement specialists so that they put Place some mitigation measures and how can we also use these automated systems in our operations if it is possible to use at least some modules in our operations to help the to prevent fraud and corruption cases in our operations and Also, uh, uh, I think that it is very important to understand that some of the modules can be used ex ante and to prevent front and corruption, but other modules can be used post factum because you need some more information to collect information from. Procurement systems to see some, for instance, collusive patterns or to see some fraud at the contract implementation stage so that some of the modules can be used by procuring entities, but other modules. Can be used by internal or external inspections and audits. So I'm sorry. I understand I don't have much time because we are approaching the end of our session, and I'm sorry that I had these issues with connection. But once again, I would like to thank the presenters and I would encourage everybody to read the report because you have much more information which is very important and which can be used in the policy dialogue, but also some of this information we can use in our operations to prevent fraud and corruption in our projects. Thank you very much. Thank you, Irina, and thank you for persevering with the technology this morning. It was great to have your, your comments at the end here. Uh, let me close the session. I think there's a couple of requests for additional references to reports, um, that, that the presenters and discussions made. So we'll make sure they go out with Natalie and the recording of the session. Uh, let me thank everybody, uh, for attending. This has been one of the most well attended, uh, governance. Uh, BBBS this year. So thank you to everyone. I think that's a good testimony to the team and their topic and their presentation that these 100 participants are still here with us at the end of a couple of hours. So thank you to you all. Thank you to those who've all been working on the report. And let me just take this moment to flag some upcoming relevant pieces of work because as Donna mentioned, we have new guidance. Coming on how to develop gov tech skills in the public sector and on the procurement side, the team have been working on a prototype for detecting uh fraud using artificial intelligence in procurement systems. So we'll be hearing more on both these topics in the future. Uh, thanks again to, to the team. Thanks to Ed for joining us this morning, and I'm wishing everybody a good evening, good rest of the day, and, uh, see you at the next one. Thank you. Thank you all. Goodbye. Thank you. All the best. Thank you so much.

showAllTranscripts

no

duration

PT1H27M52S

scene7File

worldbank/Finding_Fraud

scene7Domain

scene7FileAvs

worldbank/Finding_Fraud-AVS

title

Finding Fraud

description

Finding Fraud

showTimestampAndTranscript

yes